Emailing Design and Simulation of Four Stroke Engines

mv.
- -,-
.. 4.! * 0
j. 1.1. * * 'o
.* -
I* -
* 'A f*- 1- a
www.EngineeringBooksPdf.com
Design and Simulation
of Four-Stroke Engines
Also by Gordon P Blair:
Design and Simulation of Two-Stroke Engines

(Order No. R-161)
For more information or to-order this book, contact SAE at 400 Commonwealth Drive,
Warrendale, PA 15096-0001; (724)7764970; fax (724)776-0790; e-mail: publications ~sae.org;
web site: www.sae.org/BOOKSTORE.
Design and Simulation
of Four-Stroke Engines
Gordon P. Blair
*AM
INTfERNAIONAL
Society of Automotive Engineers, Inc.

Warrendale, Pa.
Library of Congress Cataloging-in-Publication Data
Blair, Gordon P.
Design and simulation of four-stroke engines / Gordon P. Blair.
p. cm.
Includes bibliographical references and index.
ISBN 0-7680-0440-3
1. Four-stroke cycle engines--Design and construction I. Title.
TJ790.B577 1999
621.43--dc2l 99-27316
CIP
Copyright © 1999 Society of Automotive Engineers, Inc.

400 Commonwealth Drive
Warrendale, PA 15096-0001 U.S.A.
Phone: (724)776-4841
Fax: (724)776-5760
E-mail: publications@sae.org
http://www.sae.org
ISBN 0-7680-0440-3
All rights reserved. Printed in the United States of America
Permission to photocopy for internal or personal use, or the internal or personal use of specific clients,
is granted by SAE for libraries and other users registered with the Copyright Clearance Center (CCC),
provided that the base fee of $.50 per page is paid directly to CCC, 222 Rosewood Dr., Danvers, MA
01923. Special requests should be addressed to the SAE Publications Group. 0-7680-0440-3/99-$.50.
SAE Order No. R-186
The Last Mulled Toast
A Grand Prix race is very rough,
the going's fast, the pace is tough.
The four-stroke rules the world of cars,
in bikes it's two-strokes that are the stars.
Now, why is this you'd have to ask?
The rulemakers you can take to task.
For the intake air never needs to question,
"Is this the right bellmouth for my ingestion?"
The designer ofboth must surely know,
or else his engines will all be slow,
unsteady gas dynamic trapping
by right and left waves overlapping.
To model an engine is algebraic simple.
You sit on the gas like a veritable pimple,
solving the maths the waves to track
from valve to bellmouth in the intake stack.
At the inlet valve you scan induction,
count the air that's passed by suction
and just as the valve would shut the door,
you get a wave to ram home more.
In the exhaust it's furnace hot,
for the modeller 'tis a tropic spot.
Exhaust waves reflect but do the job
of sucking out the burned gas slob.
Some time ago I wrote two tomes
on two-strokes, including poems.
It seemed only fair to tell those with cars
that black-art tuning is best kept for bars.
This book informs the four-stroke tuner
what I wish I knew those decades sooner,
as Brian Steenson followed Agostini
with my exhaust on Mick Mooney's Seeley.
The pen's both strokes have now been told.
My writ is run, I'm pensioned old.
While I may be ancient and time is shrinking,
only Dei voluntas can stop me thinking.
Gordon Blair
1 November 1998
Foreword
Since 1990, 1 have written two books on the design and simulation of two-stroke engines.
Not many in the four-stroke engine industry will read such books on the assumption that they
are not relevant to them. I will not dwell on this issue as I have already dedicated a couple of
stanzas to this very point, on the previous page. Hence, when I came to write this Foreword,
and reread what I had set down in those previous books, I realized that much of what was
written there for the two-stroke enthusiast was equally applicable to the reader of this book.
So, if much ofthis reads like the Foreword in my previous books, I can only respond by saying
that I know only one way of teaching this subject. So, if you have already absorbed that, then
pass on.
This book is intended to be an information source for those who are, or wish to be,
involved in the design of four-stroke engines. More particularly, the book is a design aid in the
areas of gas dynamics, fluid mechanics, thermodynamics, and combustion. To stop you from
instantly putting the book down in terror at this point, rest assured that the whole purpose of
this book is to provide design assistance with the actual mechanical design of an engine in
which the gas dynamics, fluid mechanics, thermodynamics, and combustion have been opti-
mized so as to provide the required performance characteristics of power, or torque, or fuel
consumption, or noise emission. Therefore, the book will attempt to explain the intricacies of,
for example, intake ramming, and then provide you with empiricisms to assist you with the
mechanical design to produce, to use the same example, better intake ramming in any engine
design. Much ofthe engine simulation, with which I was involved at QUB over the last twenty-
five years, and to which I have applied myself even more thoroughly in the three years since I
formally retired from my alma mater, has become so complex, or requires such detailed input
data, that the operator cannot see the design wood for the data trees. As a consequence, I
wound this empiricism into visual software to guide me toward a more relevant input data set
before applying it into an engine simulation computer model. Quite often, the simulation con-
firms that the empiricism, containing as it does the distilled experience ofa working lifetime,
was adequate in the first place. However, sometimes it does not and that becomes the starting
point for a more thorough design and comprehension process by simulation. You will find
many examples of that within this book. However, even that starting point is closer to a final,
optimized answer than it would have been if mere guesses had been the initial gambit for the
selection of input data to the engine simulation.
The opening of the book deals with the fundamentals of engine design and development,
ranging from mechanical principles, to engine testing and the thermodynamics of engine cycles.
To some it will read like the undergraduate text they once had; to undergraduates it will read
vii
Design and Simulation of Four-Stroke Engines
like a tutorial by some pedantic professor, and to those who had no such formal education it
will provide the thermodynamic backdrop they never had, but which they will need to follow
the logic of the design and development of the four-stroke engine. It reminds you all, expert
and novice alike, of the basics of the scene in which you wish to operate.
The acquisition of a fundamental understanding of unsteady gas dynamics is the first major
step to becoming a competent engine designer. Hence, this book contains a major section
dealing with that subject. It is little different than that within the more recent book on two-
stroke engines*, although it is updated and extended with, it is hoped, all, typographical and
theoretical errors removed. The fundamental theory of unsteady gas dynamics is the same for
two-stroke and four-stroke engines, but I repeat myself as yet another stanza has already dealt
with that. Nevertheless, without a basic understanding of unsteady gas dynamics, the myster-
ies of intake and exhaust tuning will remain just that.
The "flowing of cylinder heads" is a way of life for many developers of high-performance
engines. As with all technologies, there is a right, and a wrong, way of going about it. I explain
the only way to acquire the discharge coefficients offlow which will be meaningful if they are
also required to be accurately applied with an engine simulation model.
The discussion of combustion follows a pragmatic approach, as distinct from one steeped
in the chemistry of the subject. It provides data on the burn characteristics of a considerable
range of actual engines, spark-ignition and compression-ignition, in a manner which gives real
data input for those who wish to simulate a wide variety of power units with truly representa-
tive combustion characteristics.
The discussion on noise emission illustrates the point that actual silencers, intake, and
exhaust, can be designed by simulation so that the trade-offin noise emission and performance
characteristics can be thoroughly executed by the modeling of the entire engine together with
its mufflers. It also makes the point that the traditional empiricism, which is based in acoustics,
has a useful role to play in the design process as long as you do not believe implicitly in its
predictions.
The majority of the book is devoted to the design of the spark-ignition engine, but there is
also comprehensive treatment ofthe diesel or compression-ignition engine. The totality ofthe
book is just as applicable to the design of the diesel as it is to the gasoline engine, for the only
real difference between them is the methodology of the combustion process.
Much like this Foreword, the opening paragraphs of many of the chapters are very similar
to those in the book on two-stroke engines. I suppose it is a simple statement, albeit a truism
nevertheless, but having figured out a logical way to introduce you to the fundamentals ofany
given topic, and as those fimdamentals do not change just because I am writing about the four-
stroke engine, I decided that I would only sing the song badly if I attempted to change the
lyrics.
I have had the inestimable privilege of being around at precisely that point in history when
it became possible to unravel the technology of engine design from the unscientific black art
which had surrounded it since the time of Otto, Diesel, and Clerk. That unraveling occurred
because the digital computer permitted the programming of the fundamental unsteady gas-
dynamic theory which has been in existence since the time of Rayleigh, Kelvin, Stokes, and
*See Chapter 1, Ref. [1.9]
viii
Foreword
Taylor. For me, that interest was stimulated by a fascination with high-performance engines in
general, motorcycles in particular, and two-stroke engines even more particularly. It is a fasci-
nation that has never faded.
The marriage of these two interests, computers and racing engines, has produced this book
and the material within it. For those in this world who are of a like mind, this book should
prove to be useful.
Gordon P. Blair
23 October 1998
ix
Acknowledgements
As I explained in the Foreword, this is the third book I have written, but the
acknowledgements in those earlier books are still pertinent. The individuals who have influ-
enced my life and work are still the very same people, so what else can I say.
The first acknowledgement is to those who enthused me during my school days on the
subject of intemal combustion engines in general, and motorcycles in particular. They set me
on the road to a thoroughly satisfying research career which has never seen a hint of boredom.
The two individuals were my father, who had enthusiastically owned many motorcycles in his
youth, and Mr. Rupert Cameron, who had owned but one and had ridden it everywhere-a
1925 350 cc Rover. Ofthe two, Rupert Cameron was the greater influence, for he was a walk-
ing library of the Grand Prix races of the twenties and thirties and would talk ofengine design,
and engineering design, in the most knowledgeable manner. He was actually the senior naval
architect at Harland and Wolff's shipyard in Belfast and was responsible for the design of
some of the grandest liners ever to sail the oceans.
My father and Mr. Cameron talked frequently of two fellow Ulstermen, Joe Craig and
Walter Rusk. They are both shown in the photograph in Plate 1.0, Joe Craig standing to the
right and Walter Rusk astride the motorcycle. Walter Rusk is just about to start in the Ulster
Grand Prix of 1935 on a works Norton; he crashed on the second lap while leading. The Ulster
Grand Prix was the Grand Prix d'Europe that year, so it was an even bigger show than normal
in a province where 10% of the entire population were known, and still are known, to turn up
to watch a motorcycle race. Local media coverage ensures that racers, tuners, and engineers
are household names. Joe Craig was the Chief Engineer ofNorton Motorcycles from the early
1930s to the 1950s and he was responsible for the development of the single-cylinder 500 cc
Manx Norton throughout that period. A cutaway drawing of the 1959 version ofthis engine is
shown in Plate 1.2 and you can see its heredity etched in the lines ofthe 1935 engine in Plate
1.0. Joe Craig came from Ballymena in Co. Antrim. He designed and developed that same
engine from 31 bhp in 1931 to 53 bhp in 1953, and as a schoolboy I listened to those tales and
thought how grand it would be to know so much about tuning engines as to be able to do that.
Walter Rusk came from nearby in Whitehead in Co. Antrim and went to my alma mater, Larne
Grammar School. He was one of the top-bracket racers of the 1930s. He was killed while
flying in the Royal Air Force and I looked up at his name, written in gold on the Roll of Honor
for World War II, at school assembly every morning. These two people were my schoolboy
heroes, but ofthe two it was Joe Craig, and his genius at engine tuning, who exerted the greater
fascination.
I have to acknowledge that this book would not be written today but for the good fortune
that brought Dr. Frank Wallace (Professor at Bath University since 1965 and now retired) to
Belfast in the very year that I wished to do postgraduate research at The Queen's University of
xi
Design and Simulation ofFour-Stroke Engines
Plate 1.0 Walter Rusk and Joe Craig at the 1935 Ulster Grand Prix.
(Courtesy of Norman Windrum)
Belfast (QUB). At that time, Frank Wallace was one of perhaps a dozen people in the world
who comprehended unsteady gas dynamics, which was the subject area I already knew I had
to understand if I was ever to be a competent engine designer. However, Frank Wallace taught
me something else as well by example, and that is academic integrity. Others will judge how
well I learned either lesson.
Professor Sir Bernard Crossland deserves a special mention, for he became the Head of
the Department of Mechanical Engineering at QUB in the same year I started as a doctoral
research student. His drive and initiative set the tone for the engineering research that has
continued at QUB until the present day. I emphasize the word "engineering" because he
instilled in me, and a complete generation, that real "know-how" comes from using the best
theoretical science available, at the same time as conducting related experiments of a product
design, manufacture, build, and test nature. That he became, in later years, a Fellow of the
Royal Society, a Fellow of the Royal Academy of Engineering, and a President of the Institu-
tion of Mechanical Engineers, and was knighted, seems no more than justice.
xii
Acknowledgements
I have been very fortunate in my early education to have had teachers of mathematics who
taught me the subject not only with enthusiasm but, much more importantly, from the point of
view of application. I refer particularly to Mr. T. H. Benson at Lame Grammar School and to
Mr. Scott during my undergraduate studies at The Queen's University of Belfast. They gave
me a lifelong interest in the application of mathematics to problem solving which has never
faded.
The next acknowledgement is to those who conceived and produced the Macintosh com-
puter. Without that machine, on which I have typed this entire manuscript, drawn every figure
that is not from SAE archives, and developed all ofthe simulation software, there would be no
book. In short, the entire book, and the theoretical base for much of it, is there because the
Macintosh has such superbly integrated hardware and software allowing huge workloads to be
tackled rapidly and efficiently.
The influence of Frank Wallace and Professor Bannister turned out to be even more
profound than I had realized, for it was a reexamination of their approach to unsteady gas
dynamics that lead me to produce the engine simulation techniques described herein. Profes-
sor Bannister was the external examiner for my PhD at QUB and came from the same Univer-
sity of Birmingham which educated Frank Wallace.
I wish to acknowledge the collaboration ofall of my research students over the thirty-two
years that I worked at QUB, commencing with the late Dr. John Goulburn and concluding with
Dr. Dermot Mackey. The others will forgive me if I do not list them all-they are too numer-
ous-but any glance at the References reveals their names. Without their intellect, support,
enthusiasm, hard work, and, indeed, friendship, a great deal of that which is presented here
would be missing material.
I am indebted to those who have provided many of the photographs and drawings that
illustrate this book. Quite a few also provided experimental data, or theoretical predictions,
which are found herein. I refer to, in no particular order of precedence:
Hans Hermann of Hans Hermann Engineering
Frank Honsowetz ofNissan Motorsports
Dr. Donald Campbell of Perkins Technology
Rowland White, Norman Windrum, and Bill McLeod
Mr. Rosenthal of Classic Bike
Paul Reinke of General Motors
Melvin Cahoon of Innovation Marine
Lennarth Zander of Volvo
Douglas Hahn ofVolvo Penta
Ron Lewis of Ron Lewis Engineering
Fred Hauenstein of Mercury Marine
Dr. Barry Raghunathan of Adapco
Ing. Mario Mazuran of Seatek
Mr. Kometani and Mr. Motoyama of Yamaha Motor
Steve Wynne of Sports Motorcycles
Hau-Bing Lau (when an undergraduate), Fran9ois Drouin (when a visiting student from
the Ecole Nationale Sup6rieure des Arts et M6tiers), and Emerson Callender, Laz Foley
and Graham Mawhinney (as doctoral students) at QUB
xiii
David Holland, a QUB engineering technician, requires a special mention for the expert
production ofmany of the photographs that illustrate this book.
I cannot finish without recognizing those who helped me to establish QUB in motorcycle
road racing, for without them our design skills would have been much less evident. I refer to the
late Mick Mooney and the late Ronnie Conn of lIish Racing Motorcycles, the late Brian Steenson,
and Colin Seeley and Ray McCullough. That QUB tradition continues to this very day.
Gordon P. Blair
25 October 1998
xiv
Contents
The Last Mulled Toast ...........................................................................................................
Foreword .................................................... v
Acknowledgements ................ ...............................................
Nomenclature ............. ................................................ xi
Chapter 1 Introduction to the Four-Stroke Engine ................................................................
1.0 This Book ..................................................................................................
About
1.1 The Fundamental Method of Operation of a Simple Four-Stroke Engine .............
1.2 The Cylinder Head Geometry of Typical Spark-Ignition Engines
1.3 The Cylinder Head Geometry of Typical Compression-Ignition Engines .....1
1.4 Connecting Rod and Crankshaft Geometry .......................2
1.5 The Fundamental Geometry of the Cylinder Head ...................3
1.6 Definitions of Thermodynamic Terms Used
in Engine Design, Simulation, and Testing ........................5
1.7 Laboratory Testing of Engines ...............................7
1.8 Potential Power Output of Four-Stroke Engines ....................7
1.9 The Beginnings of Simulation of the Four-Stroke Engine ..............8
1.10 The End of the Beginning of Simulation of the Four-Stroke Engine ....... 12
References for Chapter 1 ....................................... 13
Appendix Al.1 Fundamental Thermodynamic Theory for the Closed Cycle ...... 13
Chapter 2 Gas Flow through Four-Stroke Engines ......................... 15
2.0 Introduction ..........................................15
2.1 Motion of Pressure Waves in a Pipe ...........................15
2.2 Motion of Oppositely Moving Pressure Waves in a Pipe ..............17
2.3 Friction Loss and Friction Heating during Pressure Wave Propagation .....18
2.4 Heat Transfer during Pressure Wave Pro pagation ...................19
2.5 Wave Reflections at Discontinuities in Gas Properties ................19
2.6 Reflection of Pressure Waves ...............................19
2.7 Reflection of a Pressure Wave at a Closed End in a Pipe ..............19
2.8 Reflection of a Pressure Wave at an Open End in Pipe
a ..............19
2.9 An Introduction to Reflection of Pressure Waves at a Sudden Area Change 20
2.10 Reflection of Pressure Waves at an Expansion in Pipe Area ............ 21
2.11 Reflection of Pressure Waves at a Contraction in Pipe Area ............ 21
2.12 Reflection ofWaves at a Restriction between Differing Pipe Areas ....... 21
2.13 An Introduction to Reflections of Pressure Waves at Branches in Pipes ........... 22
2.14 The Complete Solution of Reflections of Pressure Waves at Pipe Branches ..... 22
xv
2.15 Reflection of Pressure Waves in Tapered Pipes .................... 236

2.16 Reflection of Pressure Waves in Pipes for Outflow from a Cylinder ....... 239
2.17 Reflection of Pressure Waves in Pipes for Inflow to a Cylinder .......... 248
2.18 The Simulation of Engines by the Computation of Unsteady Gas Flow ..... 255
2.19 The Correlation of the GPB Finite System Simulation with Experiments ........ 288
2.20 Computation Time ...................................... 313
2.21 Concluding Remarks ..................................... 313
References for Chapter 2 ...................................... 313
Appendix A2.1 The Derivation of the Particle Velocity for Unsteady Gas Flow ........ 318
Appendix A2.2 Moving Shock Waves in Unsteady Gas Flow ............... 323
Chapter 3 Discharge Coefficients of Flow within Four-Stroke Engines ........... 327

3.0 Introduction to Discharge Coefficients .........................327
3.1 The Traditional Method for the Measurement of Discharge Coefficients . 328
3.2 The Reduction of Measured Data to Determine a Discharge Coefficient . 331
3.3 The Discharge Coefficients of Bellmouths at an Open End to a Pipe ......337
3.4 The Discharge Coefficients of a Throttled End to a Pipe ..............339
3.5 The Discharge Coefficients of a Port in the Cylinder Wall
of a Two-Stroke Engine ...................................343
3.6 The Discharge Coefficients of Poppet Valves in a Four-Stroke Engine..... 350
3.7 The Discharge Coefficients of Restrictions within Engine Ducts .........385
3.8 Using the Maps of Discharge Coefficients within an Engine Simulation . 399
3.9 Conclusions Regarding Discharge Coefficients ....................404
References for Chapter 3 ..................................... 405
Chapter 4 Combustion in Four-Stroke Engines ........................... 407
4.0 Introduction ..........................................407
4.1 The Spark-Ignition Process ................................408
4.2 Heat Released by Combustion ..............................414
4.3 Heat Availability and Heat Transfer During the Closed Cycle ...........421
4.4 Theoretical Modeling of the Closed Cycle .......................465
4.5 Squish Behavior in Engines ................................474
References for Chapter 4 ...................................... 497
Appendix A4.1 Exhaust Emissions ................................ 502
Appendix A4.2 A Simple Two-Zone Combustion Model .................. 507
Chapter 5 Computer Modeling of Four-Stroke Engines ..................... 521

5.0 Introduction ..........................................521
5.1 Structure of a Computer Model ..............................523
5.2 Physical Geometry Required for an Engine Model ..................524
5.3 Mechanical Friction Losses of Four-Stroke Engines .................534
5.4 The Thermodynamic and Gas Dynamic Engine Simulation ............537
5.5 The Ryobi 26 cm3 Hand-Held Power Tool Engine ..................537
xvi
Contents
5.6 The Matchless (Seeley) 496 cm3 Racing Motorcycle Engine ........... 547
5.7 The Ducati 955 cm3 Racing Motorcycle Engine ................... 563
5.8 The Nissan Infiniti 4000 cm3 Car Engine for the Indy Racing League ..... 574
5.9 Automobiles: A 2000 cm3 Four-Cylinder Sports-Car Engine ........... 585
5.10 Automobiles: A 2000 cm3 Four-Cylinder Turbocharged Diesel Engine ..... 598
5.11 Concluding Remarks .....................................617
References for Chapter 5 ..........................................618
Chapter 6Empirical Assistance for the Designer of Four-Stroke Engines ..........621
6.0 Introduction .......................................... 621
6.1 Empiricism for the Design of the Cylinder Head ................... 622
6.2 The Relevance of Empiricism for the Design of the Cylinder Head ................. 639
6.3 Empiricism for the Optimization of Intake System Tuning ............. 652
6.4 Empiricism for the Optimization of Exhaust System Tuning ........... 671
6.5 Concluding Remarks on Empiricism for Engine Optimization .......... 693
References for Chapter 6 .......................................694
Chapter 7 Reduction ofNoise Emission from Four-Stroke Engines ..............697
7.0 Introduction .......................................... 697
7.1 Noise ............................................... 698
7.2 Noise Sources in a Simple Four-Stroke Engine .................... 703
7.3 The Different Silencing Problems of Two-Stroke and Four-Stroke Engines ..... 704
7.4 Some Fundamentals of Silencer Design ......................... 705
7.5 Acoustic Theory for Silencer Attenuation Characteristics ............. 713
7.6 Engine Simulation to Include the Noise Characteristics ............... 729
7.7 Concluding Remarks on Noise Reduction ....................... 762
References for Chapter 7 .......................................764
Postscript-The Second Mulled Toast .................................769
Appendix-Computer Software and Engine Simulation Model .................771
Index ......................................................775
About the Author ..............................................815
xvii
Nomenclature
Most parameters are expressed in strict SI units, but custom and practice often dictate the units
below to be declared in metric but non-strict SI units. Where such units are used theoretically,
unless in the simplest of equations where the units are declared locally, they must be employed
as strict units such as m, s, N, kg, J, W, or K values.
NAME SYMBOL UNIT (SI)
Coefficients
Coefficient of heat transfer, conduction Ck W/mK
Coefficient of heat transfer, convection Ch W/m2K
Coefficient of heat transfer, radiation Cr W/m2K4
Coefficient of friction Cf
Coefficient of discharge Cd
Coefficient of discharge, actual Cda
Coefficient of discharge, ideal Cdj
Coefficient of contraction c
Coefficient of velocity Cs
Coefficient of loss of pressure, etc. CL
Squish area ratio Csq
Coefficient of combustion equilibrium Kp
Area ratio of engine port to engine duct k
Manifold to port area ratio Cm
Local port area ratio Ct
Valve acceleration ratio 9v
Valve ramp lift ratio Cr
Valve lift ratio Lr
Modified valve lift-diameter ratio LD
Intake tuning ramming factor Cir
Exhaust tuning factor for primary pipe length Cet
Exhaust tuning factor for collector tailpipe length C
Exhaust collector pipe area ratio Cco
Dimensions and physical quantities
area A m2
diameter d m
length x m
xix

length of computation mesh L m
mass m kg
molecular weight M kg/kgmol
radius r m
time t s
volume V m3
force F N
pressure p Pa
pressure ratio P
pressure amplitude ratio X
mass flow rate m kg/s
volume flow rate m3/s
velocity of gas particle c m/s
velocity ofpressure wave propagation a m/s
velocity of acoustic wave (sound) a m/s
Young's modulus Y N/m2
wall shear stress T N/M2
gravitational acceleration g m/s2
Dimensionlessnubr
Froude number Fr
Grashof number Gr
Mach number M
Nusselt number Nu
Prandtl number Pr
Reynolds number Re
Energy Mwork anld heat related 1aaetr
system energy E 3
specific system energy e J/kg
internal energy U 3
specific internal energy u J/kg
specific molal internal energy U J/kgmol
potential energy PE 3
specific potential energy pe J/kg
kinetic energy KE 3
specific kinetic energy ke J/kg
heat Q 3
specific heat q J/kg
enthalpy H 3
specific enthalpy h J/kg
xx
Nomenclature

specific molal enthalpy h J/kgmol
entropy S J/K
specific entropy s J/kgK
work W J
work, specific w J/kg
Engin e. physical agorney
number of cylinders n
cylinder bore dbo m
piston area Abo m2
cylinder stroke Lst mm
bore to stroke ratio Cbs
connecting rod length Lcr mm
crank throw Lct mm
swept volume vsv m3
swept volume, trapped Vts m3
clearance volume vcv m3
compression ratio, geometric CR
compression ratio, trapped CRt
speed of rotation N rev/min
speed of rotation rpm rev/min
speed of rotation rps rev/s
speed of rotation C) rad/s
mean piston speed cp mA/
crankshaft position at top dead center tdc degrees
crankshaft position at bottom dead center bdc degrees
crankshaft angle before top dead center °btdc degrees
crankshaft angle after top dead center °atdc degrees
crankshaft angle before bottom dead center °bbdc degrees
crankshaft angle after bottom dead center °abdc degrees
crankshaft angle 0 degrees
angle of obliquity ofthe connecting rod degrees
combustion period bo degrees
throttle area ratio Cs
exhaust blowdown time-area W s/im
exhaust pumping time-area Qep s/m
exhaust overlap time-area *eo s/m
intake ramming time-area *ir s/m
intake pumping time-area Vjp s/m
intake overlap time-area C s/m
intake valve opens ivo degrees
xxi

intake valve closes ivc degrees
exhaust valve opens evo degrees
exhaust valve closes evc degrees
Engefe. cierforems related parameters
mean effective pressure, brake bmep Pa
mean effective pressure, indicated imep Pa
mean effective pressure, friction fmep Pa
mean effective pressure, pumping pmep Pa
power output W kW
power output, brake Yb kW
power output, indicated Wi kW
torque output Z Nm
torque output, brake Zb Nm
torque output, indicated Zi Nm
air-to-fuel ratio AFR
air-to-fuel ratio, stoichiometric AFRS
air-to-fuel ratio, trapped AFRt
equivalence ratio A
equivalence ratio, molecular Xm
specific emissions of hydrocarbons bsHC g/kWh
specific emissions of oxides of nitrogen bsNOX g/kWh
specific emissions of carbon monoxide bsCO g/kWh
specific emissions of carbon dioxide bsCO2 g/kWh
specific fuel consumption, brake bsfc kg/kWh
specific fuel consumption, indicated isfc kg/kWh
air flow, scavenge ratio SR
air flow, delivery ratio DR
air flow, volumetric efficiency N
charging efficiency CE
trapping efficiency TE
scavenging efficiency SE
thermal efficiency Th
thermal efficiency, brake lb
thermal efficiency, indicated ms
mechanical efficiency TIM
fiuel calorific value (lower) Cfl MJ/kg
fuel calorific value (higher)
fuel latent heat of vaporization
Cfhi MJ/kg
hvap UJ/kg
mass fraction burned B
heat release rate QRq J/deg
xxii
Nomenclature

combustion efficiency TiC
relative combustion efficiency wTt purity 'nse
relative combustion efficiency wrt fiueling Tlaf
index of compression ne
index of expansion nc
flame velocity Cfl
flame velocity, laminar Cjf
flame velocity, turbulent Chb
squish velocity CS
engine speed for intake ramming peaks Nrp rev/min
engine speed for intake ramming troughs Nrt rev/min
Gas properties
gas constant R J/kgK
universal gas constant R J/kgmolK
density r kg/m3
specific volume v m3/kg
specific heat at constant volume Cv J/kgK
specific heat at constant pressure Cp J/kgK
molal specific heat at constant volume Cv J/kgmolK
molal specific heat at constant pressure Cp J/kgmolK
ratio of specific heats y
purity HI
temperature T K
viscosity A kg/ms
kinematic viscosity v m2/s
volumetric ratio of a gas mixture u
mass ratio of a gas mixture
Noise
sound pressure level dB
sound intensity I W/m2
sound frequency f Hz
attenuation or transmission loss ,tr dB
wave length of sound A m
perforation opacity ratio 0
General
vectors and coordinates x, y, z
differential prefixes,
exact, inexact, partial and incremental d, 8, f, A
xxiii
Chapter 1
Introduction to the Four-Stroke Engine

1.0 About This Book
It is generally accepted that the theoretical cycle on which the four-stroke engine is based
was proposed by Beau de Rochas in 1876. The fist practical demonstration ofthe engine was
implemented by Otto in 1876. This book is not about the history of the internal-combustion
engine, but realizing that some of you may wish to study it, it is recommended that you peruse
the informed writings of Cummins, Obert, Taylor, Caunter, or Ricardo [1.1-1.5]. The book by
Cummins [1.1] is quite an authoritative text in this historical context.
This book is also not about the detailed design of the mechanical components of an
engine, such as crankshafts or connecting rods. For that, one reads elsewhere in the literature.
Nor is it a comprehensive collection of design ideas for the cylinder head, valving, or ducting
geometries of every configuration of four-stroke engine constructed in times past.
This book is about the design of the four-stroke engine so as to achieve its target perfor-
mance characteristics for the application required, irrespective of whether that application is
intended for Formula 1 car racing or a lawnmower. To do that, one must thoroughly under-
stand the filling and emptying of the engine cylinders with air and exhaust gas and the
combustion of the trapped charge within them. Hence, this book is about the unsteady gas
dynamics and thermodynamics associated with the four-stroke engine. Nevertheless, to sensi-
bly design for the performance characteristics, one must bring the real geometry of the engine,
its cylinder head, combustion chamber, mifolding, and ducting into the gas dynamic and
thermodynamic design process, otherwise the outcome is meaningless, not to mention useless.
Therefore, very frequently, the real geometry and the measured test data from actual engines
will be produced to illustrate a design point being made. To conduct such a design process, the
only pragmatic approach is to simulate the unsteady gas dynamics and thermodynamics within
the entire engine, basing the simulation on the physical geometry of that engine in the finest
detail, from the aperture where air enters the engine initially to the aperture where the exhaust
gas finally exits from the engine.
1.1 Fundamental Method of Operation of a Simple Four-Stroke Engine
1.1.1 The Four Strokes That Make Up the Cycle
The simple four-stroke engine is shown in Fig. 1.1, with the several phases of the filling
and emptying of the cylinder illustrated as Figs. 1. l(A)-(D). The sketch shows the engine as a
spark-ignited unit with access to the cylinder controlled by poppet valves actuated by cams,
(A) INTAKE STROKE (B) COMPRESSION STROKE (C) POWER STROKE (D) EXHAUST STROKE
Fig. 1.1 The four strokes of the four-stroke engine.
tappets, and valve springs. The mechanical definition of some, if not all, of such words
describing salient engine components is shown sketched on Fig. 1.2. The cylinder contains a
piston with sealing effected by two piston rings and a third lower ring, which scrapes the
excess lubricating oil off the cylinder walls back into the crankcase, which acts as an oil sump.
The movement of the piston is controlled by a rotating crankshaft with the connecting rod
linking the small-end bearing on the piston, through a gudgeon pin, to the big-end bearing on
the crankpin. The word "gudgeon pin" is a British word; in the United States it is referred to as
a "wrist pin." A flywheel is attached to the crankshaft to help smooth out the torque pulsations.
In the sketch, Fig. 1.1, it is seen as a bob-weight flywheel.
Plate 1.1 shows a cross section through an engine much like that sketched in Fig. 1.1. All
of the items described briefly above can be seen in this photograph, but it provides a realistic
scaling of the many components. The intake tract is on the right of the picture and the exhaust
duct is to the left.
It is clear from Fig.. 1.1 that the maximum movement of the piston within the cylinder, i.e.,
the stroke of the piston, is, simplistically, twice the upward, or downward, movement of the
length ofthe crank. The crank length is often referred to as the "crank throw." The maximum
movement of the piston creates a volume swept by the piston, i.e., the swept volume, Vsv, and
it is clear from the sketch that the piston stops short of the cylinder head face, thereby creating
Chapter I - Introduction to the Four-Stroke Engine
EXHAUST
Fig. 1.2 An overhead valve engine with two valves per cylinder.
a minimum volume space in which the combustion of trapped air and fuel can take place. This
minimum volume is defined as the clearance volume, Vcv. When the piston is at the top or
minimum cylinder volume position, it is referred to as being at top dead center, tdc, and when
it is at the bottom or maximum cylinder volume position it is described as being at bottom
dead center, bdc. In certain books, mostly historical, tdc and bdc are sometimes called inner
and outer dead center, idc and odc; they will not be referred to as such within this book. This
volumetric behavior gives rise to the concept of a compression ratio for the engine. For the
four-stroke engine this ratio is always employed as the geometric compression ratio, CR:
CR = maximum volume = V1d - Vsv + Vcv (1.1.1)

minimum volume Vtdc VCV
The employment of a significant level of charge compression within an engine like this
was the major inventive contribution of Otto [1.1].
Plate 1.1 A cross section through a two-valve spark-ignition engine.
1.1.2 The Four Strokes

Air is induced into this engine by an intake stroke of the piston from tdc to bdc, thereby
increasing the cylinder volume from its minimum to its maximum value. This is shown in
Fig. 1.1(A). The intake poppet valve is actuated by the movement of the rotating cam depress-
ing the tappet and spring. Ideally, it opens at tdc, lifts to a maximum value at about mid-stroke,
and ideally closes at bdc. The effect is somewhat similar to when you open your mouth, sharply
intaking air into your lungs, and then close your mouth again; such an intake process is caused
by your lungs expanding and air entering because the atmosphere is then at a higher pressure
than within your lungs. Thus, as the piston moves from tdc to bdc on the intake stroke, the
cylinder fills-ideally with a mass of air equivalent to a swept volume at atmospheric pressure
and temperature. Whether this air brings with it the requisite fuel for future combustion, either
by a carburetor placed within the intake duct, or by a similarly placed fuel injector squirting
fuel into the air as it passes by, or by a fuel injector that sprays fuel directly into the air now
within the cylinder, is really a matter of implementation for any particular design
configuration.
Chapter I - Introducton to the Four-Stroke Engine
The air now trapped within the sealed cylinder experiences a compression stroke as the
piston moves from bdc to tdc. The volume decreases from maximum to minimum, i.e., from
Vbdc to Vtdc. The pressure and temperature of the air rises and the trapped fuel associated with
this air vaporizes. This is illustrated in Fig. 1.1(B).
At the conclusion of the compression stroke, combustion takes place, initiated by an elec-
tric spark from the spark plug shown located between the valves in the cylinder head in Fig.
1.1. The pressure rises rapidly to many tens of atmospheres, and the temperature soars to
several thousand degrees Celsius. Ideally, all of this takes place at tdc and instantaneously.
The piston now descends on the power stroke, pushed by the very high pressure difference
across it from the upper side facing the cylinder to the lower side seeing the crankcase at
atmospheric pressure. The cylinder volume goes from Vtdc to Vbdc. This force, i.e., a push, on
the piston gives a torque on the crank, which in turn powers whatever the engine is driving.
This is sketched in Fig. 1.1(C).
Because the cylinder now contains the products of combustion, i.e., an exhaust gas, which
must be removed if ever the cylinder is to breathe air again and become a cyclic device, an
exhaust stroke is initiated as shown in Fig. 1.1(D). The exhaust poppet valve is actuated by a
well-timed turn of the rotating cam, which depresses the tappet and compresses the valve
spring and lifts the valve. The exhaust valve ideally opens at bdc, lifts to a maximum value at
about mid-stroke, and ideally shuts again at tdc. The exhaust gas is forcibly expelled by the
upward sweeping movement ofthe piston through the aperture created by the annulus around
the head of the lifted exhaust valve. The exhaust stroke occurs when the piston moves from
bdc to tdc and the cylinder volume decreases from maximum to minimum, i.e., from Vb& to
Vtdc. Using the human body again as an analogy, the exhaust stroke is somewhat similar to you
filling your lungs with air and opening your mouth, thereby letting your lungs decrease in
volume with the exhalation; you then close your mouth again. That exhaust process is caused
by your lungs deflating, and your personal exhaust gas exiting, because the atmosphere is at a
lower pressure than within your lungs.
1.1.3 Analogy between the Human Body and an Engine
As I have often told my students, the analogy between the human body and the internal
combustion engine is quite uncanny. We breathe in air, and breathe out exhaust gas, by varying
the volume of our lungs. We consume fuel and some of us actually produce useful work! We
are even the epitome ofthat thermodynamic ideal, expert practitioners of the isothermal heat
transfer of the Carnot cycle [1.2]. We characterize the power output of the internal combustion
engine by the horsepower, derived by a direct comparison with another animal engine. There
are even exhaust emission analogies to be explored, but this text will not descend any further
down this rich vein of humor, already infamously over-exploited during my academic teach-
ing career to a ribald undergraduate audience. On a more serious note, and in this very context,
only a century ago the Sanitation Officer of the City of New York, then populated by tens of
thousands of horses, declared that he was looking forward to the day when the noxious odors,
not to mention particulate emissions, producing health risks associated with these animals,
would be eliminated by the arrival of the "horseless carriage and its clean engine." A hundred
years later that debate has been tumed on its head, with the intemal combustion engine now
relegated to the role of polluter.
1.2 Cylinder Head Geometry of Typical Spark-Ignition Engines

1.2.1 Two-Valve Overhead Valve Engine
The most common four-stroke engine produced is arguably the overhead valve (ohv) engine
with two valves per cylinder. This type of engine is sketched in Fig. 1.2, with an elevation
section through the valves at top; underneath is a plan view looking at the cylinder head face.
The sketch shows the actuation of the valves by two overhead camshafts, colloquially
referred to as a double overhead camshaft engine, and often written simply as a dohc design.
Plate 1. 1 shows a cross section through just such an engine. Many famous racing engines were
built just like this. The 350 or 500 cm3 Manx Norton racing motorcycles were outstanding
examples in their own era of the high specific power output that could be extracted from this
type of power unit; the engineering talents of the Chief Engineer of Norton, Joe Craig [1.35],
have already been referred to in the Foreword. A drawing of the 1959 model of the 500cc
Manx Norton engine is shown in Plate 1.2 with the overhead camshafts driven by a vertical
shaft through bevel gear drives, and with the valve motion controlled by hairpin valve springs.
Plate 1.2 The 500cc Manx Norton racing motorcycle engine. (Courtesy of Classic Bike)
Chapter 1 - Introduction to the Four-Stroke Engine
As already stated, although this book is not a compendium ofthe myriad ingenious meth-
ods that the engineer has designed to actuate the poppet valves in an engine, I propose to
illustrate some relevant cylinder head geometry so that the discussion elsewhere, such as in
Chapter 5, is more meaningful. Historically, the most popular method to actuate the two
valves is the pushrod and rocker arm system, again often simply referred to as a dohv layout.
The basic mechanism of its operation is shown in Fig. 1.3. The camshaft is driven from the
engine crankshaft and pushes on a tappet, which lifts the pushrod, oscillates the rocker arm,
and lifts and lowers the valve off, and back on to, the valve seat. An adjuster is employed to
control the clearances of the elements of the mechanism. This is shown here as a simple
mechanical device, but human ingenuity has yielded tappets that are sliding hollow cylinders
filled with oil which can control the pushrod end clearances automatically; these are known as
hydraulic tappets. Pushrod designs have ranged from solid steel rods to hollow aluminium
tubes. Clearly, the higher is the location of the camshaft, the shorter is the pushrod and the less
it will be inclined to bend or buckle when pushed by the tappet.
Fig. 1.3 Valve actuation by a pushrod and rocker arm.
A cutaway drawing of a two-valve ohv engine with pushrod actuation is shown in Plate
1.3. This is of the 500cc Gold Star BSA, which was used in both on-road and off-road racing
throughout the world, and was particularly popular in the United States in the era when Daytona
Beach racing was actually held on the beach! There was an on-road version using this engine,
as a sports-touring machine, which must have been one of the most handsome motorcycles
ever built. As a student at QUB in the 1950s, I virtually turned green with envy just looking at
them through the showroom window of W.J. Chambers & Co. in Donegall Pass in Belfast.
Other options include a single overhead camshaft (sohc) moving rocker arms directly onto
each valve that eliminates the need for pushrods (see Plate 5.6 in Chapter 5). References 1.1-
1.8 include excellent drawings, sketches, and photographs of many engines developed through-
out history and their methods of valve actuation.
Plate 1.3 The 500cc Gold Star BSA racing motorcycle engine.
(Courtesy of Classic Bike)
Irrespective of how the valve is actuated, a typical valve-lift crank angle graph is as shown
in Fig. 1.4, where the exhaust valve opens and closes at points marked as evo and evc, respec-
tively. The intake valve opens and closes at points marked as ivo and ivc, respectively. The
x-axis is crankshaft (or crank) angle. On this particular graph, this angle is zero at tdc on the
power stroke, and lasts for 7200; the durations of the individual strokes are clearly marked.
Irrespective of the method ofvalve actuation, several fundamental points emerge from an
examination of Figs. 1.2 and 1.4. The first point to be observed is that the valves do not, and
cannot, lift instantaneously. They have mass, i.e., inertia, when accelerated, spring forces must
be overcome, and friction resistance requires that force must be exerted to effect movement.
All ofthis simply means that a real valve takes time to lift and requires work input to do it. The
idealized set ofstrokes described in Sec. 1.1.2 assumes simplistically that the valves opened or
closed instantaneously, at tdc or bdc, as the need arose. Fig. 1.4 makes clear the reality that an
exhaust valve will typically open some 110 °atdc, i.e., 700 before bdc, and will close some 30
°atdc and not at tdc. All of this is necessary to ensure that the valve exposes sufficient aperture
area to the cylinder so as to release the combustion products from it. It would be possible to
drop the valve more rapidly back onto its seat at tdc. However, the ensuing bouncing behavior
of what is, in effect, a spring-mass system would ensure that it would actually stay open for
even longer and would incur such large impact forces on the cam and tappet as to seriously
reduce its durability from excess wear and erosion. Nevertheless, the less the mass of this
entire oscillating system, the higher is the rotational speed it can reliably run. Thus, in racing
TlC BDC TDC BDC TPC

power exhaust '| intake compression,
stroke stroke stroke stroke
'EXHAUST: INTAKE
IVALVE VALVE
, OVERLAP
PERIOD
w E I
"CRANK EVO IVO EVC IVO
Fig. 1.4 Typical valve liftprofiles with respect to crank angle.
engines it is not surprising to find double overhead camshaft systems employed. For lower-
speed engines, it becomes possible to withstand the extra mass and inertia provided by rocker
arns and pushrods, as shown in Fig. 1.3, without significant loss ofvalve lift or timing error at
the maximum speed of operation.
The second point to be observed is that the thermodynamic cycle occupies four strokes of
the piston whereas the total crankshaft angle it covers is 7200, i.e., two complete rotations of
the crankshaft. However, because each valve is operated over the 3600 period ofthe intake and
exhaust strokes, and then must remain dormant during the 360° period of the two strokes that
make up the compression and power strokes, it follows that the camshaft is driven at one-half
the speed of rotation ofthe engine crankshaft. In short, as the crankshaft drives the camshaft,
typically by either chain, toothed timing belt, or gears, the gearing ratio for this drive between
crankshaft and camshaft is 0.5, i.e., the camshaft turns at half the engine speed.
The third point to be observed from Fig. 1.4 is that there is a period between the exhaust
and intake strokes when the intake valve is opening and the exhaust valve is closing, i.e., both
valves are open together. This is called the valve overlap period. The potential for mechanical
or gas flow mayhem is obvious. Apart from the valves actually colliding, ifthe cylinder pres-
sure is too high as the intake valve opens, then large quantities of exhaust gas can be shuttled
into the intake tract. This gas is hot, maybe 1000°C, and can cause anything from backfire to
coking-up fuel residues on the back of the intake valve. At the very least, every particle of
exhaust gas sent up the intake tract will occupy space normally reserved for air and, when real
cylinder inflow eventually occurs, only a reduced mass ofair will be induced into the cylinder.
As the air mass induced on each cycle ultimately equates to power attained, this can hardly be
described as an optimum breathing procedure. On the other hand, if at the same juncture the
exhaust pressure is lower than in either the intake tract or the cylinder, then the intake process
can begin early on the intake stroke, so there exists the possibility of enhancing the mass of air
inhaled. Overdoing this effect can actually send fuel-laced air into the exhaust pipe, which will
ultimately appear in the atmosphere as an excess of unburned hydrocarbon emissions. Hence,
based on these preliminary remarks, the design process for the engine during the valve overlap
period is surely seen as a critical affair. But I get ahead of myselfbecause this and many other
similar subjects, such as the effectiveness of intake ramming behind a still-open intake valve
on the compression stroke, are really what the rest of this book is all about!
The two-valve engine type, as sketched in Fig. 1.2, appears in many formats as can be seen
in the literature [1.1-1.8]. Some have the inclined valves and a part-spherical combustion cham-
ber, essentially as illustrated here. Such an engine is called a "hemi-head" engine, from the
word "hemisphere," which rather describes the shape of the combustion area. Some have the
valves vertical, as in the diesel, or compression-ignition, engines to be discussed later. Hence,
most diesel engines have the combustion bowl within the piston crown.
1.2.2 Two-Valve Side Valve Engine
The simplest of all four-stroke engines is the side-valve engine, shown sketched in
Fig. 1.5.
10
Fig. 1.5 A side-valve engine with two valves per cylinder.
Many a car in my youth used nothing else. Some were famous, such as the Ford V8 Pilot
in the saloon car, and the 1172 cm3 Ford IOOE four-cylinder engine that powered many a
single-seat racing car in Britain and Ireland in the fifties and sixties, and is the forerunner of
that used in today's very popular Formula Ford racing. The car in which I learned to drive, my
father's Morris Series E, had a 1000 cm3 side-valve engine. Being possessed ofrather indif-
ferent suspension and steering, it was perhaps fortunate that its modest power output limited
its top speed to a near-lethal 50 mph!
Fig. 1.5 shows that the valves are indeed at the "side" of the cylinders and, due to the
physical limitations imposed by the bore dimension and nornal inter-cylinder spacing, it is
virtually impossible to employ other than one exhaust valve and one intake valve per cylinder.
Moreover, because these limitations control the maximum size of the valves that may be
employed, the breathing ability of the engine is compromised. The valve design is quite rigid,
11
in that the camshaft is near the crankshaft. Indeed, in many ways it is as effective as an over-
head camshaft engine design in this regard. The cylinder head is as simple as a loop-scavenged
two-stroke engine and the total design is not only compact and light, but also very economical
to manufacture. This is why tens of millions of these were made by Briggs and Stratton of
Milwaukee for industrial, agricultural, and horticultural applications. The major design prob-
lem is that all gas flow into, or out of, the engine is very restricted by the throat between the
valve chest area and the engine cylinder, causing pumping work losses in both directions.
Upon the compression stroke, further pumping is required ofthe piston to get the fresh charge
back over the valve chest area to be ignited by the spark plug. The combustion chamber shape
is poor and prone to detonation, so a relatively low compression ratio must be used. Thus, heat
transfer, combustion, and pumping work losses are high in this engine, which reduces its power
output and fuel economy and raises its output of most of the legislated exhaust-gas emissions.
1.2.3 Four-Valve Pent-Roof Cylinder Head
This design for spark-ignition (si) engines has become very common in automobiles and
motorcycles in the last decade of the twentieth century. Such an engine is shown sketched in
Fig. 1.6 as a dohc layout and even that further sophistication is to be found in a high proportion
of automobiles and motorcycles today. At one time, the four-valve dohc engine was reserved
only for racing engines, but first the Japanese motorcycle industry, then the Japanese car
industry, and then the rest of the world, employed it universally in general automotive produc-
tion.
There have been many four-valve engines used throughout history. A Rudge motorcycle
won the Ulster Grand Prix in 1928 and 1929, and the company subsequently produced a four-
valve road-going equivalent of the Very successful racer, and named it the Rudge "Ulster"
model. My father owned one before the start ofWorld War II and I can dimly remember him
going to and fro in his work on this elegant black-and-gold machine. A cutaway drawing ofthe
1938 version ofthis engine is shown in Plate 1.4, where the valve actuation is by pushrods and
rocker-arms. The intake valves are parallel, much as in a modern pent-roofhead layout, but the
exhaust valves are rather more radially disposed. The piston crown has a shallow dome.
A comparison of Figs. 1.2 and 1.6 shows that it is possible to get more valve head area in
the four-valve design and so aid the engine to breathe more air during induction. The so-called
pent-roof layout means that the valves are disposed on the sides of a wedge, or roof-like,
combustion chamber with the added possibility of squish areas disposed around the cylinder.
The spark plug can be located centrally, which gives equal flame travel paths to the remote
corners of the combustion chamber. Although in the two-valve engine shown in Fig. 1.2 the
spark plug are shown to be centrally located, it is more common to use larger valves than those
sketched there, and to locate the plug toward one side of the axis of the two valves.
The 4v dohc design is now commonplace in current automobiles and the cutaway drawing
of a 1998 2.4 liter I4 car engine by General Motors, shown in Plate 1.5, perfectly illustrates the
genre. The camshaft drive is by a roller chain.
12
- - - - - - - - - - 0- - - -
Fig. 1.6 A four-valve pent-roof cylinder head.
1.2.4 Porting the Cylinder Head by Other Means

There have been designs for porting the cylinder head by means other than the use of
poppet valves. The first of real note is the use of sleeve valves which rotate-oscillate is a
better word-around the engine cylinder and expose holes in the cylinder wall for the intake
and exhaust processes, much as is seen for the porting of a two-stroke cycle engine. A cylinder
wall port, almost by definition, opens more rapidly and exposes more aperture area to the
cylinder than can a poppet valve [1.6, 1.9]. Hence, the engine has the possibility of breathing
more air and expelling its exhaust gas more easily. Caunter [1.4] describes the Barr and Stroud
and other, similar, sleeve valve engines at the turn of the twentieth century, and Ricardo [1.6]
shows excellent scale drawings and photographs ofmany of these designs. A cutaway drawing
of the 1922 350cc Barr and Stroud motorcycle engine is shown in Plate 1.6. The valve is
known as the Burt-McCollom type and its motion is a combination of reciprocation and semi-
13
Plate 1.4 The 500cc Rudge Ulster motorcycle engine.

rotation, i.e., as in a figure eight. Its mechanical reliability, because it was designed in the
period before Sir Harry Ricardo established some logical design principles for sleeve-valve
engines, was not outstanding.
Perhaps the most notable ofthese sleeve-valve designs, to operate on gasoline with spark-
ignition, were the Bristol Hercules and Centaurus aircraft engines. They were far removed
from simple designs, mechanically speaking, as the Hercules had a 14-cylinder, air-cooled,
radial layout.
The second and different type of valving is the rotary valve installed in the cylinder head.
Here, the poppet valves are replaced by a rotating valve which spins concentrically with the
cylinder axis, or a barrel valve where the spin axis is at right angles to it. R.C. Cross of Bath
was one of the originators ofresearch into this type of engine [1.10], and I was greatly privi-
leged to hear him deliver his Chairman's Address to the Automobile Division ofthe Institution
of Mechanical Engineers, in Belfast in 1958, when I was a student. It was a brilliant lecture
14
Plate 1.5 A 2.4 liter in-linefour-cylinder automobile engine. (Courtesy of General Motors)
but, as I listened to it quite enthralled, it certainly never occurred to me that I would follow him
into his Chairman's role some thirty years later! The book by Hunter [1.11] provides an impor-
tant reference for the geometries of virtually every rotary-valve engine ever built.
Due to the higher values of exposed breathing area, all of these engines, sleeve-valve or
rotary-valve, in their day did deliver higher power outputs than the equivalent engines fitted
with poppet valves. Their Achilles' heel was the exposure to the cylinder, and more impor-
tantly to its combustion chamber, of the rotating surfaces of the valves which are, however
lightly, smeared with lubricating oil. The result was the partial burning ofthat oil, which raised
the oil consumption rate of the engine, gave a somewhat smoky exhaust containing particu-
lates, and tended to coke up the valve, thereby not only reducing its durability but also
decreasing the breathing area of the aperture.
Wankel Engines
I do not intend to discuss the Wankel, or rotary-piston four-stroke engine within the pages
of this book. Firstly, it is such a special case that it would distort every descriptive section of
this book to the point where the debate on the conventional reciprocating engine would be lost
in the caveats needed to cope with the Wankel engine. Secondly, as I write this book in 1998,
I cannot think of a single Wankel engine that is in current production. So, there appears to be
15
Plate 1. 6 The 350cc Barr and Stroud sleeve-valve motorcycle engine.

no current technical interest in the engine or its development. If there is such a Wankel engine
in current production, then I apologize, but it has not come to my attention. Thirdly, I can well
imagine why the Wankel is not in current production for automobiles and the automotive
industry as a whole, but I can think of several areas where it ought to be seriously reconsidered
for sound technical reasons. However, even to tell you about that would require a decent-sized
book all by itself. Hence, the word "Wankel" will not reappear within these pages.
1.3 Cylinder Head Geometry of Typical Compression-Ignition Engines
1.3.1 General
This engine was invented by Rudolph Diesel in 1892 as a means of burning coal dust. He
found out rather quickly that coal dust was inferior to liquid fuels for a compression-ignition
combustion process. The engine carries his name to this day, being colloquially referred to as
a "diesel." One argument holds that Ackroyd-Stuart in England produced an engine of this
16
type before Diesel. Be that as it may, today Rudolph Diesel conventionally gets the credit.
Combustion in a compression-ignition (ci) engine occurs by using a compression ratio that is
sufficiently high so as to produce an in-cylinder air temperature such that the vapor surround-
ing the injected fuel droplets is heated to its self-ignition temperature. Whereas the spark-
ignition, using the more volatile gasoline fuel, has compression ratios between 8 and 11, the
compression-ignition unit requires compression ratios that are typically between 18 and 21 to
accomplish its combusfion process. To prevent such a combustion process from being too vigor-
ous, a fuel that contains heavier hydrocarbons, and is less volatile, than gasoline is employed.
Thus, the spark-ignition engine uses a gasoline, which is typically based on octane of the
paraffin family, with a chemical composition of C8H18, whereas the compression-ignition
engine uses a fuel based on dodecane ofthe same paraffin family, with a chemical formula of
C12H26. This subject is discussed in m6re detail, in Sec. 1.6.6 ofthis chapter, and in Chapter 4.
The basic process is one of heating air by compression, then using the hot air to vaporize
a liquid fuel and heat the vapor to its self-ignition temperature. Such heating processes take
time, so it is not surprising to find that diesel engines tend to run more slowly than spark-
ignition engines. The basic mechanism that can be used to speed the process up is to raise the
heat transfer rate between the hot air and the cold fuel droplets by increasing the relative
velocity between these two fluids. This presents two options: (1) either have fuel droplets that
move rapidly through the air, or (2) have slower-moving fuel be whirled by faster-moving air.
The first option gives rise to the direct-injection diesel engine and the second to the indirect-
injection type.
1.3.2 The Four-Valve Direct-Injection (DI) Engine
The necessity of having a high compression ratio is discussed briefly above. Because the
needed compression ratio is typically about twice that of the gasoline engine, it follows that
the clearance volume of the combustion chamber is about half that of the spark-ignition
engine. It becomes almost impossible to achieve a such a small clearance volume and, at the
same time, arrange the intake and exhaust valves any way other than vertical with respect to
the cylinder axis. The layout of a four-valve engine is shown in Fig. 1.7 and the vertical dispo-
sition of the valves is clear. Because the piston runs very close to the head face at tdc, it
becomes very difficult to achieve the small clearance volume except by having the minimum
ofvalve overlap periods. With any other arrangement, the valves will qontact the piston crown.
A high-pressure fuel injector is seen in the middle ofthe four valves and in the middle of
the cylinder, squirting fuel in sprays of fast-moving fine droplets toward the edge of a combus-
tion chamber, which is normally located as a bowl in the piston. To further enhance the process
of heating the fuel, the air may be made to swirl around the vertical axis, albeit slowly by
comparison with the indirect-injection engine discussed below in Sec. 1.3.3. The swirl ratio,
i.e., the speed of rotation of this air vortex, is rarely more than five times higher than the engine
speed. The in-cylinder air swirl is created during the induction process by suitably orienting
the incoming air-direction past the intake valves; you may glance ahead at Plate 3.3. The
spinning vortex is then somewhat accelerated by being compressed into the combustion cham-
ber toward the end of compression. The fuel injection line pressure is typically in the range of
500 to 1200 atn, with the start of injection normally some 15 °btdc.
17
Fig. 1. 7A four-valve direct-injection (DI) diesel engine.
The particular shape of chamber sketched in Fig. 1.7 is known as, for obvious reasons, a
"Mexican hat." Once again, engineering inventiveness has produced a plethora of designs for
diesel combustion and you are referred to Lilly [1.12] for a compendium of their geometries.
On a historical and personal note, this book is the successor to that founded by C.C. Pounder,
who was the Chief Engineer of Harland and Wolff's shipyard in Belfast and a famous designer
of marine diesel engines. Note, too, that fuel injection technology for both gasoline and diesel
fuel will be discussed later in this book but, if you cannot wait, then Lilly may again be con-
sulted [1.12].
The design shown in Fig. 1.7 has four valves per cylinder. Many designs throughout his-
tory had but two valves. The reason that four-valve designs have become more common is that
diesel engines for automobiles have gained great popularity, particularly in Europe, because of
good fuel economy and, in recent years by using turbo-chargers, good power output as well.
The four-valve layout permits better air breathing and lower pumping losses, particularly
18
considering the necessary minimum nature of the valve overlap period. Modem, turbo-charged
DI engines for automobiles, with total capacities typically between 2 and 3 liters, now run to
over 4000 rpm, and have specific outputs at around 55 kW per liter. For trucks, or marine
pleasure craft, the cylinder size is usually about 1 liter per cylinder, but the engine speeds are
lower and normally do not exceed 3000 rpm.
A cutaway drawing of a turbocharged truck diesel engine is shown in Plate 1.7. It is an 16
design, a "straight-six" in the jargon of the automotive world, with two valves per cylinder
actuated by pushrods. The turbocharger can be seen at the left. The swept volume per cylinder
ofthis engine is about I liter. The combustion bowl within the piston, for direct injection of the
fuel, are clearly drawn as are the vertical disposition of the valves. A closer look at one ofthe
combustion bowls in this engine is provided by the photograph in Plate 1.8. This design is
known as a "squish lip" with the bowl being considerably reentrant. The compressed air is
swirled and squished from the cylinder head area into the combustion bowl; evidence of these
effects can be found in Sec. 4.5. Note also the large bearing areas at the small-end and the
considerable length, bulk, and, inevitably, mass of this piston compared to that for the spark-
ignition engine seen in Plate 1.1.
Plate 1. 7 A cutaway drawing ofa turbocharged direct-injection six-cylinder diesel engine.

(Courtesy ofPerkins Engines)
19
Plate 1.8 A cutaway drawing of the piston for a direct-injection diesel engine.
(Courtesy ofPerkins Engines)
1.3.3 The Three-Valve Indirect-Injection (IDI) Engine

This engine design typifies the other fuel-heating option of high-speed in-cylinder air-
flow and the slower moving fuel created by lower pressure fuel injection systems. Here, the
fuel injection line pressure is nornally in the range of 150 to 300 atm. The basic layout of the
engine type is shown in Fig. 1.8.
Upon compression, the air is pumped into a side chamber and here the shape sketched
mimics that developed by Ricardo [1.6] and is known as a Comet chamber. This is described
as indirect injection, abbreviated as IDI, and is still widely employed in European diesel auto-
mobiles. The swirling air flow within the Comet chamber rotates at up to 25 times the engine
speed. Because the engine can run to a higher speed than the DI design, i.e., up to 4500 rpm,
this means that this air vortex spins at up to 90,000 rpm. The result is a "cleaner" combustion
process than the DI engine, i.e., cleaner in terms of exhaust emissions.
20
Fig. 1.8 A three-valve indirect-injection (IDI) diesel engine.
There is a price to pay for this smoother, and quieter-running, IDI diesel engine. The
pumping process through the small orifice connecting the cylinder and the side chamber is
paid for in pumping work put into it during compression. There are furither losses on expansion
as the combustion charge blows down through the same orifice to the main cylinder during the
power stroke. The orifice-to-piston area ratio is quite severe, i.e., the area of the connecting
orifice is only about 1% of the bore area. For example, ifthe cylinder bore is 90 mm, then the
orifice diameter would be about 9 mm. The result ofthe energy consumed during these pump-
ing processes is that there is about a 10% diminution of power and a similar diminution of the
fuel efficiency of the engine.
As the air flows into the side chamber through the orifice upon compression, the air tem-
perature falls in correspondence with the pressure drop. This air temperature decrease is con-
siderable, which has implications for both normal running and cold start-up. The result is that
the IDI engine requires a compression ratio about one ratio higher than the DI unit, and
requires a glow plug to get the combustion going in a cold-start situation.
21
The design shown in Fig. 1.8 has three valves per cylinder, typically seen in recent PSA
(Peugeot-Citroen) automobiles, and is probably the optimum cylinder head layout for this
type of engine. Nevertheless, many cylinder head configurations have been proposed for the
IDI engine type. Lilly [1.12] should be consulted for more diagrammatic design detail.
1.3.4 Exhaust Emissions
The diesel engine, like its spark-ignition counterpart, is under legislative pressure to con-
form to ever-tighter emissions standards. Even though it provides very low emissions of
carbon monoxide and of hydrocarbons, the diesel engine does emit visible smoke in the form
of carbon particulates and measurable levels of nitrogen oxides, the latter being considerably
lower for the IDI than the DI engine. The level of emissions of both of these power units is
under increasing environmental scrutiny and the diesel engine must conform to more stringent
legislative standards by the year 2000. The combination of very low particulate and NOX
emission is a tough R&D proposition for the designer ofdiesel engines to be able to meet. As
the combustion is lean of the stoichiometric mixture by some 50% at its richest setting in order
to avoid excessive exhaust smoke, the exhaust gas is oxygen rich and so only a lean burn
catalyst can be used on either a two-stroke or a four-stroke diesel engine. This does little, if
anything at all, to reduce the nitrogen oxide emissions. This subject is covered more com-
pletely in Chapter 4.
1.4 Connecting Rod and Crankshaft Geometry
We have reached the point now where some mathematical treatment of design will begin.
This will be conducted in a manner that can be followed by anyone with a mathematics educa-
tion of university entrance level.
1.4.1 Units Used throughout this Book
Before embarking on this section, a word about units is essential. This book is written in
SI units, and all mathematical equations are formulated in those units. Thus, all subsequent
equations are intended to be used with the arithmetic values inserted for the symbols of the SI
units. Units are listed in the Nomenclature section or within the text. If this practice is fol-
lowed, then the value computed from any equation will appear also as the strict SI unit listed
for that variable on the left-hand side of the equation. Should the user desire to change the unit
of the ensuing arithmetic answer to one of the other units listed in the Nomenclature section, a
simple arithmetic conversion process can be easily accomplished. One of the virtues of the SI
system is that strict adherence to those units, in mathematical or computational procedures,
greatly reduces the potential for arithmetic errors. I write this with some feeling as one who
was educated with great difficulty, as one of my American friends once expressed it so suc-
cinctly, in the British "furlong, hundredweight, fortnight" system of units!
22
Chapter I - Introduction to the FourStroke Engine
1.4.2 Swept Volume

If the cylinder of an engine has a bore, dbo, and a stroke, Lst, as sketched in Fig. 1.9, then
the swept volume, VSV, of the cylinder is given by the product of cylinder bore area, Abo, and
stroke length:
Abo 4
dbo VSV = Abo x Lst = 4 dboL (1.4.1)
The total swept volume, Vtsv, often referred to simply as the "capacity," of an engine with
a number of cylinders, n, is found from:
Vtsv = nVsv = n 4 dboLst (1.4.2)
1.4.3 Compression Ratio

The geometrical compression ratio, CR, is already defined in Eq. 1.1.1, but this equation
can be manipulated so that the clearance volume, VCV, may be calculated from the swept vol-
ume of any cylinder and its compression ratio, thus:
VCV CR-i (1.4.3)
Theoretically, the actual compression process occurs after the intake valve closes, and the
compression ratio after that point can be important in design terms. The volume swept by the
piston from intake valve closure to tdc is referred to as the trapped swept volume, Vts. This
defines a trapped compression ratio, CRt, which is then calculated from:
CRt = vts + VCV (1.4.4)

VCV
1.4.4 Piston Position with Respect to CrankshaftAngle
The piston sweeps up and down the cylinder displacing volume and, in any thermody-
namic simulation of an engine, it is vital to know the precise cylinder volume at any crankshaft
angular position. The piston is connected to the crankshaft by a connecting rod of length Lcr
as seen in Fig. 1.9.
The throw of the crank is simplistically one half of the stroke under most normal circum-
stances, but is correctly designated as length Lct. As with two-stroke engines, the ratios of
connecting rod to crank throw are typically in the range between 3 and 4.
23
lesign and Simulation of Four-Stroke Engines
Fig. 1.9 The geometry of a connecting rod-crank mechanism.
At any given crankshaft angle, 0, after the tdc position of the crank, the connecting rod
enter line assumes an angle + to the cylinder center line. This angle is often referred to in the
terature as the "angle of obliquity" of the connecting rod. This is illustrated in Fig. 1.9. The
earing location where the connecting rod attaches to the piston is called the "small end" and
iay be offset by an amount D to the cylinder center line. The same geometrical effect is given
y the small-end bearing being located centrally on the piston, but the cylinder axis is offset by
n amount D from the crankshaft center line. Clearly, it is possible to accommodate numeri-
ally any totality of pin and/or cylinder axis offset. Note that the offset value, D, is positive if
iat offset is toward the direction of crank rotation; this happens to be the situation shown in
ig. 1.9.
When the crank is located at its tdc angular position, i.e., when 0 is zero, it is clear from the
ietch that the piston may not be at its tdc position, i.e., its furthermost extension up the
ylinder. If there is any pin or cylinder axis offset, i.e., if D is not zero, it is clear that this
annot be the case and the tdc and bdc piston positions will occur at Otdc and bdc, respectively.
24
It is also clear that Otdc and Obdc are not equal. Using the nomenclature of the sketch it is
possible to solve for the following unknown variables.
From the center panel of Fig. 1.9, with the piston at its maximum extension:
Ftdc + Gtdc= (Lr + Lt)-D2 (1.4.5)
hence 0tdc= tan'(D (1.4.6)
hence Gtdc= Lct cosOtdc (1.4.7)
and Ftdc Lct cos Otdc + (Lcr + L2c)-t (1.4.8)
From the bottom panel of Fig. 1.9, with the piston at its minimum extension:
Fbdc = |(Lcr- ct)D2 (1.4.9)
hence Obdc tan-( (1.4.10)
and G bdc Lct (1.4.11)
The stroke of the piston is then given by:
Lst = Ftdc + Gtdc - Fbdc (1.4.12)

If the gudgeon pin offset and/or cylinder axis offset is zero, i.e., if D is zero, then Eqs.
1.4.5-1.4.11, reduce to:
Otdc = ° Obdc = 0
Ftdc Lcr + Lct

=
Gtdc Lct =
25
Fbdc = Lcr - Lct

and the length of the stroke becomes, from Eq. 1.4.12:
LSt =
Ftdc + Gtdc Fbdc = Lcr + Lct - (Lcr -Lct) = 2Lct
- (1.4.13)
Consider the position within the cylinder of any point on a piston with respect to its
motion from its tdc position to a point where the crank has turned through an angle 0 from the
tdc angular position of the crank. For convenience, a point, marked as X, is located at the
small-end bearing center and its location down the cylinder from its tdc position is Ht. This is
shown sketched on the upper part of Fig. 1.9.
The controlling trigonometric equations are, in solving for the length Ht:
Ht = (Ftdc + Gtdc) - (F + G) (1.4.14)

where E=Lct sinO (1.4.15)
where G= Lct cosO (1.4.16)
and F= Lr-(E-D)2 (1.4.17)
Consequently, using the information in Eq. 1.4.5 and Eqs. 1.4.14-1.4.17, the piston posi-
tion is:
Ht = (L2r + L Dt)D2 - cr - (L sin0 - D)2 - Lct cos 0 (1.4.18)
The angle of obliquity ofthe connecting rod, ¢, is given by:
ta E DJ)LCtsinO-DJ
= - (1.4.19)
The angles employed during, or determined during, the analysis of the above equations
are in radians. To convert any angle from radians to degrees, or vice-versa, it is useful to
remember that 3600, i.e., a fill circle, is equivalent to 2In radians.
26
A more complex trigonometrical question is often asked of the designer or modeller. If the
distance from tdc on the piston is measured as Ht, then at what angle 0 has the crank turned
from its tdc angular position? From Eq. 1.4.14, where k becomes a known constant for any
given data set:
F = (Ftdc + Gtdc-Ht )-G = k-G (1.4.20)

then k=Ftdc +Gtdc-Ht
Squaring this equation, and inserting F and G from Eqs. 1.4.15-17 yields:
asin0+bcos0+c = 0 (1.4.21)
where a = -2DLct b = 2kLct c = k2 L_ct + D (1.4.22)
thenlet t = tana where a=-

2
consequently sine 2t and cos0= -

I +t2 I +t2
Substitution of the values related to t into Eq. 1.4.21 reduces it to a quadratic equation,
with t being the unknown variable:
(c - b)t2 + (2a)t + (b + c) = 0 (1.4.23)
The solution for the crank angle, 0, becomes
0 =2 tanf1I -abc +b- )(1.4.24)
Because the values of a, b, c, and k are all known, the solution for the crank angle, 0, at any
piston position, Ht, can be determined directly.
27
1.4.5 Numerical Data on Piston Position With Respect to CrankshaftAngle

Sinusoidal Motion Compared to Connecting Rod-Crank Motion
Firstly, let us consider a simple numerical example of an engine with a bore dimension of
86 mm, a crank throw of 43.mm, and a connecting rod of length 165 mm. Let us also consider,
in this first example, that the gudgeon pin offset, D, is zero. In that case, the piston position at
its tdc position coincides with the crank at its tdc angular position. The above equations are
solved and the piston position is plotted in Fig. 1.10, as a percentage of the maximum stroke
which, by definition, is always a percentage of the swept volume. It can be seen, in Fig. 1.10,
that the maximum stroke occurs at 180' crank angle. The stroke is 86 mm and the swept
volume is 499.557 cm3. The connecting rod obliquity is calculated from Eq. 1.4.19 as 15.110
and occurs at a crank angle of 900, and again at 2700, where it is -15.11°.
The motion of the piston is seen to be almost sinusoidal, which would be the solution for
piston position in Eq. 1.4.18 ifthe connecting rod were infinitely long, in which case the value
of Ht would be simply LctcosO. This true sinusoidal motion is also plotted in Fig. 1.10 and it is
quite clear that intolerable stroke and volumetric errors would occur if simple sinusoidal
motion were to be assumed for the motion of the piston. The magnitude of that error is plotted
in Fig. 1.11 and the maximum value of it is seen to occur at crank angle positions of 900 and
2700 with a value close to 7%.
Effect ofGudgeon Pin Offset
Secondly, let us consider the same numerical example but with gudgeon pin offsets, D, of
+2 mm and -2 mm.
If D is +2 mm, the stroke ofthe engine, Lt, becomes 86.007 mm as calculated using Eq.
1.4.12, and the swept volume is now 499.597 cm3 from Eq. 1.4.1. The piston tdc positions are
no longer at 00 and 1800 on the crank rotation, but at 0.55 1 and 180.9390, respectively. These
angles correspond to those calculated as Otdc and Ob& in Eqs. 1.4.6 and 1.4.10. Maximum
connecting rod obliquities still occur at 900 and 2700 crank angle, but the values are now
14.390 and -15.830, respectively.
If D is -2 mm, the stroke of the engine, Lst, remains at 86.007 mm as does the swept
volume at 499.597 cm3. The piston tdc positions are also no longer at 00 and 1800 on the crank
rotation, but are now at 359.450 and 179.060, respectively. Maximum connecting rod obliqui-
ties still occur at 900 and 2700 crank angle, but the values are now changed to 15.830 and
- 14.390, respectively.
Fig. 1.12 plots these differences in piston position with respect to a zero pin offset design,
as a finction of crank angle. The stroke error is shown as a percentage. It can be seen that
stroke error is maximized at about 900 and 2700 crank angle, with maximum values of ±0.6%.
During engine simulation, because cylinder volume can only be computed from piston
position as a function of crank angle, if the gudgeon pin is offset, then the volumetric error,
typified by the 0.6% calculated above, cannot be tolerated because this would give rise to
computed pressure and temperature errors of at least the same magnitude. During any analysis
of measured cylinder pressure diagrams, the error in the crank angle location of the piston
28
CONNECTING ROD-CRANK
100
LL
80
0-
Lli
60
a:
0
z 40
C'
0-
20
0
0 90 180 270 360
CRANKSHAFT ANGLE, Qatdc
Fig. 1.10 Sinusoidal and connecting rod-crank motion.
ERROR OF CONNECTING ROD-CRANK MOTION

WITH ZERO PIN OFFSET
7 COMPARED TO SINUSOIDAL MOTION
6
LLF
0I- 5
a:
CO) 4
z
0 3
a:
0
c: 2
a:
w
1
0
0 90 180 270 360
CRANKSHAFT ANGLE, 2atdc
Fig. 1.11 The error by assuming sinusoidal motion for a connecting rod and crank.
29
0.7
0.6
0.5
-0.4
y 0.3
0.2
0
0
-0.2
-0.3
wr -0.4
CO -0.5
g 1. The
2 PINcofgudeo
E pin o PetOFFpstoEmtin
-0.6
0 90 180 270 360
CRANKSHAFT ANGLE, Oatdc
Fig. 1.12 The effect ofgudgeon pin offset on piston motion.
position at tdc is known to be very significant [1.13] and will have a major bearing on the
accuracy of calculation of engine indicated mean effective pressure and heat release. But I get
ahead of myself again, for these matters are discussed later in Chapter 4.
It should be clear from the above that the engine modeller has no option other than to
simulate accurately the geometry ofthe crankshaft and the piston motion using the theory set
out in Eqs. 1.4.1-24. Anything less, or the use of assumptions or approximations, leads sim-
plistically to a case of "gigo" syndrome, i.e., "garbage in is garbage out"!
1.5 The Fundamental Geometry of the Cylinder Head
The cylinder head of a four-stroke engine normally contains intake and exhaust poppet
valves. As they lift, they expose area between the cylinder and the connecting duct, either an
intake or an exhaust pipe, and gas will flow into, or out of, the cylinder depending on the
prevailing pressure difference between them. During the simulation of an engine it is essential
to be able to calculate these geometrical areas at any juncture in the rotation ofthe crankshaft.
There are two aspects to this requirement. The first is the exposed valve area at any particular
valve lift, and the second is the valve lift characteristic with respect to crank angle. This sec-
tion is dedicated to solving these twin requirements for simulation purposes.
If the engine is a sleeve valve, or a rotary valve, engine you are referred to my book on
two-stroke engines [1.9], where the debate on the aperture areas of ports at a cylinder is con-
ducted in great detail, and the relevant calculation procedures are described.
30
1.5.1 Connecting the Valve Apertures with the Manifold

That four-stroke engines normally contain poppet valves within the cylinder head can be
seen in Figs. 1.1-1.8. Parametric detail regarding the valve aperture and connecting duct areas
is observed in Fig. 1.13.
In any given cylinder head, the number of valves for either the intake or the exhaust is
designated as n, and the valves are theoretically distinguished by the further subscript appella-
tion of an 'i' or an 'e,' so the number of intake and exhaust valves is niv and nev, respectively.
Similarly, as the general nomenclature for a port is Ap, the flow area at the port for each of
these valves is Aip and Aep, respectively. If there is more than one intake, or more than one
exhaust, valve then the total exposed port area is Aipt and Aept, respectively. Normally, the
valve(s) connect to a single manifold outlet of area Am, which gives rise to the nomenclature
of Aem and Aim for the exhaust and intake manifold areas, respectively. This gives rise to the
concept of pipe (manifold)-to-port area ratios, Cm, which are defined as follows for the
exhaust and intake ducts in the cylinder head:
Intake: im n
niv x Aim
xAi ip Aipt(15) (1.5.1)
Exhaust: Cem = n A (1.5.2)

nev x ep Aept
PIPE TO PORT AREA RATIOS, Cem & Cim

Cim - A.imC'emAem
A.pt
INTAKE EX-HAUS
Alim - i - Awn
Aipt ilv Ap / ,=v

n Aep
p pistonr
Fig. 1. 13 Thie pipe-to-port area ratios for the fou'r-stroke engine.
31
As will be shown in later chapters, these manifold-to-port ratios are critical for the perfor-
mance of an engine for this area ratio, as either an expansion or a contraction, controls the
amplitude of any pressure wave created within the ducting by the cylinder state conditions.
Because the strength of any such unsteady gas flow tuning is a function of its pressure, the
connection becomes obvious.
The bottom line of Eqs. 1.5.1 and 1.5.2 contains a term as yet not clearly defined, namely
the throat or minimum port aperture area at any valving, Apt. This is shown sketched on Fig.
1.14.
Inflow or outflow at any valve passes through the valve curtain areas which correspond to
the side areas of a frustum of a cone. However, at the highest valve lifts, the minimum flow
area is normally that at the inner port where the diameter is dcl,, which is further reduced by the
presence of a valve stem of diameter d". These are the controlling aperture areas for the
exhaust and intake ports at the valves, which are defined as follows:
Aept = nevAep =nev (di- d2) (1.5.3)
Aipt = nivAip = ni (d - d2) (1.5.4)
Should any port not be reduced to dip from the inner seat diameter, d3,, then the value of dis
should replace djp in Eq. 1.5.3 or 1.5.4.
1.5.2 The Geomney of the Aperture Posed by a Poppet Valve
The physical geometry of the poppet valve and its location, as shown in Fig. 1.14, is
characterized by a lift, L, above a seat at an angle +, which has inner and outer diameters di.
and dos, respectively. The valve curtain area, At, for this particular geometry is often simplis-
tically, and quite incorrectly, expressed as the side surface area of a cylinder of diameter di-
and height L as:
At = idisL (1.5.5)
It is clear from the sketch that it is vital to calculate correctly the geometrical throat area of
this restriction, At. In Fig. 1.14, the valve curtain area at the throat, when the valve lift is L, is
that which is represented by the frustum of a cone defined by the side length dimension, x, the
32
Fig. 1.14 Valve curtain areas at low and high valve lifts.
valve seat angle, 4, the inner or outer seat diameters, dis and dos, and the radius, r, all ofwhich
depend on the amount of valve lift, L. The side surface area of a frustum of a cone, A., is:
As (major +dminor J1X (1.5.6)
where x is the length of the sloping side and dmnmor and dmajor are its top and bottom diameters.
This is the maximum geometrical gas flow area through the seat of the valve for flow to, or
from, the pipe beyond the port, where that minimum port area is Aipt or Aept, as defined above.
The dimension x through which the gas flows has two values, which are sketched in Fig.
1.14. On the left, the lift is sufficiently small that the value x is at right angles to the valve seat
33
and, on the right, the valve has lifted beyond a lift limit, Llim, where the value x is no longer
normal to the valve seat at angle ¢. By simple geometry, this limiting value of lift is given by:
Llim
Liim = dos dis
2s
-
= dos -dis
2(5 (1.5.7)
2 sin cos l sin
For the first stage of poppet valve lift where:
L < Llim
the valve curtain area, At, is given from the values of x and r as:
x= LcosO (1.5.8)
r = dis + x sin p (1.5.9)

2
in which case At = nLcos4(dis + Lsinicos4) (1.5.10)
For the second stage of poppet valve lift where:
L > Llim
the valve curtain area, At, is given from the higher value of x as:
X= -L OstanJ
-
.( Os2 (1.5.11)
whence At = Osd + Ld-J

j(L - d i tan.J + (dOs dis (1.5.12)
If the seat angle is 450, which is conventional, then tan(O) is unity and Eq. 1.5.12 simpli-
fies somewhat.
34
Thus, the total exposed annular flow areas of the intake and exhaust valves, Ait and Aet,
are given at any particular valve lift by:
Aet = nev(At)exhaust (1.5.13)
Ait = niv(A)inlet (1.5.14)
A simple examination of these equations, and particularly the inaccurate approximation of

Eq. 1.5.5, reveals that lifting any valve by more than about one quarter of its inner seat diam-
eter would appear to be paradoxical, as the geometrical areas of the throat and port would
become equal at that juncture. In practice, valves are lifted much higher to about 0.35 or 0.4 of
the value of the inner seat diameter. The reason is that the effective area of flow through the
valve seat is reduced by fluid mechanic flow losses, colloquially referred to as, and numeri-
cally encapsulated as, a discharge coefficient, Cd [1.16-1.18].
Considering the situation for a single valve layout, the flow "fills" the port or minimum
area where that area is Ap. At a conventionally high valve lift, the valve throat area, At, nor-
mally exceeds that of the port area, Ap. If a discharge coefficient is defined as Cd, and will be
seen in Chapter 3 to be measured for the totality of the port and valve throat restriction, then
the effective area of that restriction becomes Ate, which is defined as follows:
At < Ap Ate =Cd x At (1.5.15)
The topic of measurement and application of discharge coefficients is discussed at length

in Chapter 3. It is not a trivial subject, but one of great importance, and has come under consid-
erable reexamination in recent times [1.16-1.18].
To ensure accuracy regarding the prediction of mass flow rates and of the magnitude of
pressure wave formation, the above analysis of the valve curtain area must be used in any
simple empirical analysis and, equally importantly, within any competent engine simulation
code, and in the derivation of the measured discharge coefficients [1.16-1.18].
1.5.3 The Lift Characteristics of a Poppet Valve
It can be seen in Fig. 1.4 that a poppet valve can neither be lifted nor dropped instanta-
neously. The several elements of a typical valve lift diagram are shown in Fig. 1.15.
The valve commences to lift at a crank angle vo, and upon closing returns to zero lift at
crank angle vc. In this context, when examining manufacturers' specifications or technical
papers [1.34] for valve lift, and the opening and closing crank angle locations, it is often found
that these opening and closing points are quoted at some nominal lift value such as 0.5 or 1.0
mm. I have never heard a satisfactory technical explanation of why the valve timings and lifts
35
::
IIg
' IZ A
a L
a.ANiz : 0
a.
1< I <
cr.
o~ VC
Fig. 1.15 Valve lift characteristics as afunction of crank angle.
are quoted in this manner. Because this information is useless to the modeller, who must know
the precise static timings of all valves so as to compute the valve aperture areas at any instant
during crankshaft rotation, great care must be talcen to ensure that the "gigo" syndrome has not
developed.
A poppet valve lift is to be found in some five differing phases, as seen in Fig. 1. 15. IThe
maximum lift of the valve is Lv and the total duration is Ov° crank angle.
IThe first phase is the opening ramp, designed to lift the valve gently offits seat; perhaps
that should be restated as being as rapidly as one dares at the highest speed of engine opera-
tion. This phase lasts for a cranlc angle period Of Our during which the valve lifts from zero to
a height of Lu,,r In any discussion on valve lift, the conventional units for crank angle are in
degrees and valve lift in mm dimensions. Note also that angular periods are quoted as crank
rotation, and not cam rotation which is always one-half that ofthe crankshaft.
The second phase is the main lift from the end of the first ramp to the start of the dwell
period around the maximum liftc point. During this crank angle period of Our degrees, the valve
lifts from Lur to Lv.
The third phase is a dwell around peak lift when the valve remains at Lv for period Of Od,
crank angle degrees.
Tvhe fourth phase is the valve drop from the end of dwell to the commencement ofthe final
ramp. This lasts for Od, degrees and the valve falls from a lift of Lv to Ldr.
The fifth and final phase is the ramp to valve closing, which lasts for Odr degrees, with the
valve falling from a lift of Ldr to zero.
36
In real engines, it is common practice to have the opening and closing ramps be identical
in duration and to have similar values for Lus and Ldr. The common practice in spark-ignition
engines is to have values for ramp duration at 400 and, for such a duration, to have the valve lift
by some 20% of the maximum lift value during that period. The common practice in (lower-
speed) compression-ignition engines is for the ramp lift to be as high as 50% ofthe maximum
lift. There is no common practice regarding the duration of the dwell period, but 50 is consid-
ered to be a fairly common value, although a dwell period is sometimes absent.
1.5.4 The Acceleration and Velocity Characteristics of a Poppet Valve
At any point during the lift, or drop, of a poppet valve, let us assume that the valve move-
ment is dL during a time period dt when the engine speed is N rpm. The engine will rotate for
a crank angle period of dO during this time interval. Time and crank angle are related in the
normal manner:
dt =
60/N _ 115.6
dO 360 6N (1.5.16)
The velocity of the valve during this interval, cV, is given by, if the lift is quoted in mm
units:
1 dL 1 dL dO 6N X-
dL
cv=~-
1000 dt
~x-x-----
1000 dO dt 1000 dO
mJs (1.5.17)
If the velocity ofthe valve changes by an amount dcv during the time interval dt then the
acceleration or deceleration of the valve, gv, is given by:
dc
=-x-= -x-xX-= x-d
1 cI
g dO 1 6N dc
Xvdt g dO dt g 9.81 dOg (1.5.18)
where g is the acceleration due to gravity, which has a value of 9.81 m/s2.
In practical terms, one has either measured, or has designed, a valve lift characteristic that
reduces to a data file, which normally consists ofvalve lift, Lv, in mm, at crank angle intervals,
dO, which are typically in steps of about one degree. Consider the situation sketched in Fig.
1.16.
Fig. 1.16 consists of three points on a typical valve lift diagram where the lifts are L1, L2,
and L3 at crank angle positions 01, 02, and 03, respectively. Using the theory in Eq. 1.5.17 to
determine the mean velocity in these two elements, cvi and cv2:
6N L2 - L1 6N L3 - L2
0=0-X
1000 02 - (I CV2 -m/ 03 - 02
1000 (1.5.19)
(..9
37
1 02 03
/
*
E
E
i
Li
CRANK ANGLE, 09
Fig. 1.16 Three adjacent points on a valve lift curve.
The mean acceleration, gvl2, for the lift process as motion from the median point of the
first element, to the median point on the second element, is given by Eq. 1.5.18 as:
6N CV2 -cvl (1.5.20)

g *81 005{(02+03) - (01 +02)}
If it happens, as would be normal, that all crank angle intervals on the valve lift data file
are identical, and have a value AO, then Eqs. 1.5.19-1.5.20 reduce to:
6N
- x L2 L1
- 6N L3 - L2 6N cV2 -
Cvl
cVI =
1000 AO cv2 = 1000
xo AO
AH gv12 = -8 x
9.81 AO
(1.5.21)
Thus, one can move from point to point on a known valve lift curve and find the velocity
and acceleration characteristics at each point upon it, at any given speed of engine operation.
1.5.5 Designing the Valve Lift Characteristics of a Poppet Valve
During the modelling and simulation of an existing engie, the engineer will be certain to
have available the measured lift characteristics of the valves in the actual engine. However, if
the engine does not exist, the modeller has to be able to rapidly create a realistic valve lift file.
Even the process of optimization of an existing engine will almost certainly include extending
that optimization to its valving and valve events. So, here also arises the need for the creation
38
of a different, yet realistic, set of valve lift files. The word "realistic" is used to describe a
proposed valve lift file that would permit the design and manufacture of the rest of the valve
train, i.e., cams, springs, tappets, pushrods, etc., without there being excessive accelerations or
velocities on the several elements within it. The traditional route is to design a new cam based
on, for example, polydyne principles, including the rest of the valve train in what is a very
complex mathematical exercise in mechanics [1.18-1.21]. The modeller, being by instinct a
gas dynamicist or a thermodynamicist, does not have the time, nor indeed the training, to carry
out this exercise into a theoretical specialism all by itself. One could use a specific CAD
design software package if it is readily available, but be assured that even the use of such
software involves a very considerable amount of training.
I decided a long time ago that a simpler, yet effective and realistic, method for the rapid
creation of a new valve lift file was essential [1.22]. An examination of the valve lift files from
many engines showed that, if they were reduced to specific lift L., at specific angles Os, they
exhibited very similar characteristics. Specific lift and specific angle, Ls and Os, are defined in
terms of the nomenclature seen in Fig. 1.15, where the lift is L( at an angle 0 from the com-
mencement of lift, and where the maximum lift and valve opening duration are Lv and Ov,
respectively:
Ls LO (1.5.22)
Os = (1.5.23)
Further examination revealed that it would be extremely difficult, if not impossible, to

formulate a single mathematical expression to accurately describe a specific valve lift curve.
Furthermore, it would be very limiting in its coverage ofall possible valve lift characteristics,
even if it could be accomplished. It is essential to be able to design separately, in a mathemati-
cal sense, valve lifts that have differing up and down ramps, and varying amounts of dwell.
Thus, the approach typified in Fig. 1.17 was adopted.
Fig. 1.17 shows that specific lift and specific angle relationships are described separately
for a ramp period and a lift period. The same polynomial relationship is used for the "ramp
down" as is used for the "ramp up," and similarly for the "main lift up" and the "main lift
down." The relationship linking specific lift and angle is a third-order polynomial in each
case, the coefficients of which are determined from an analysis of measured data. The func-
tions are as follows, as shown in Fig. 1.17:
For ramp up or down: Ls = kro + kri0s + kr20S + kr30s (1.5.24)
For main lift up or down: Ls = km0 + kmiOs + km20S + km30s (1.5.25)
39
Ls=kmo+kmijs+km20s2+km3S3 LS=1
MAIN LIFT Ls=0

Os=0 OS=1
LS=1
o Ls=kro+kri 0s+kr0s2+k3Os3
o
w
RAM LS=O
cn OS=0 OS=1
SPECIFIC ANGLE, Os
Fig. 1.17 Specific lift characteristics of a poppet valve.
The procedure for simulating any valve lift curve is to decide on the following basic
parameters. The nomenclature here is as shown in Fig. 1.15. First, the opening and closing
angles of the valve are "designed," i.e., numerical values are assigned to vo and vc, as crank
angles atdc, where zero crank angle is tdc on the power stroke and extends for 7200. The valve
opening duration, Ov, is then given by:
0V =vc-vo (1.5.26)
The duration of the dwell angle, Odw, is "designed" as a number of crank angle degrees,
typically in the range of 0 to 10. The ramp-up and ramp-down periods, 0r and Od, are
"designed," i.e., each is assigned a number of crank angle degrees. Similarly, the main lift
periods, up and down, Ou1 and Odl, are each assigned a number of crank angle degrees. The
totality ofthese angular values must add up to Ov, as seen in Fig. 1. 15. Thus, the creation ofthe
valve lift diagram, in total period or element duration terms, is completely flexible.
40
However, custom and practice, not to speak of simplicity of cam profile manufacture,
dictates that the ramp periods, and the main lift periods, are normally identical. Hence, using
this design process, I usually set the main lift periods to be identical and I normally use 400 for
each ramp period. There are two reasons for this: (1) I have rarely found it necessary to do
otherwise, and (2) all my polynomial coefficients listed below were reduced from measured
data over a 400 crank angle ramp period. Hence:
O°u =0dl = ev -dw 2 our Odr (1.5.27)
The next "design" decision to be made is the amount of lift to be assigned to the up and
down ramps, i.e., Lur and Ldr. As stated earlier, as a proportion of the maximum lift, Lv, these
numbers are rarely less than 20% for si engines or greater than 50% for ci units. Although there
is no numerical problem in assigning different values to the "up" and "down" ramp lift ratios,
symmetry of a valve lift profile is much more common in lift characteristics than asymmetry.
Asymmetry, normally in the form of a higher lift ratio for the "ramp down" to the "ramp up,"
is sometimes seen in high-performance racing engines, particularly in Formula 1 engines. The
lift ratios for the ramps, up and down, are defined as C. and Cdr, respectively.
Ramp lift ratios: Cur = Cdr = Ldr (1.5.28)
With these "design" decisions made, the valve lift curve can be computed by moving from
element to element in sequence, as shown in Fig. 1.17, starting with valve opening and the
opening "ramp up."
The Valve Lift Commences
At a crank angle 0 of zero:
0=0 L0=0 (1.5.29)

The Opening Ramp Up
At any angle 0, where:
0 < 0 < our (1.5.30)
0
Specific angle: Os = (1.5.31)
Our
41
Insert Os into Eq. 1.5.24 and calculate Ls. Note that the opening point, where 0 is zero, is
not calculated here, but is positively declared as being zero above. The value ofactual lift, Le,
is found by translating specific lift into actual lift at the crank angle 0:
Actual valve lift: Le = CurLSLV (1.5.32)
The Main Lift Up

Our0 <Our + Oul (1.5.33)
Specific angle: Os =
0-Our (1.5.34)
Oul
Insert Os into Eq. 1.5.25 and calculate Ls. The value of actual lift, L9, is found by translat-
ing specific lift into actual lift at the crank angle 0:
Actual valve lift: Lo = Lur + Ls(Lv - Lur) (1.5.35)
The Dwell Period

our + Oul < 0 < our + Oul + (dw (1.5.36)
Specific angle: es = 1 (1.5.37)

The value of actual lift, L0, is simply the maximum lift:
Actual valve lift: Lo = Lv (1.5.38)

The Main Lift Down
our + Oul + (dw <0 < our + Oul + (dw + (dl (1.5.39)
Specific angle: 0s = our + Oul + Odw + Odl -0 (1.5.40)

Odl
42
Chapter I - Introducdon to the Four-Stroke Engine
It should be noted that the values of angle and lift are determined by their position from the
start of the down ramp, i.e., the computation is operated in reverse for valve drop by compari-
son with valve lift. Insert Os into Eq. 1.5.25 and calculate Ls. The value of acual lift, Le, is
found by translating specific lift into actual lift at the crank angle 0:
Actual valve lift: Lo = Ldr + Ls(Lv - Ldr) (1.5.41)
The Ramp Down

our + Oul + Odw + Odl < 0 < Ov (1.5.42)
Specific angle: 0 = oV -e (1.5.43)

Odr
It should be noted that the values of angle and lift are determined by their position from the
start of the down ramp, i.e., the computation is operated in reverse for "ramp down" by com-
parison with "ramp up." Note also that the final point atOv is not calculated, but is reserved for
a positive "shut" in the next segment below. Insert Os into Eq. 1.5.24 and calculate Ls. The
value of actual lift, Lo, is found by translating specific lift into actual lift at the crank angle 0:
Actual valve lift: Lo = CdrLsLv (1.5.44)
The Valve Shuts

At crank angle 0:
O =OV LO = O (1.5.45)
The use of positive zeroing of the valve lift curve, at opening and closing, takes care of the
numeric problems caused by the polynomial coefficient k,0 in Eq. 1.5.24 not being an actual
zero when the measured data for valve lift ramps are "best fitted' with a third-order polyno-
mial function.
Smoothing the Computed Valve Lift Curve
Upon combining the several elements of a valve ift curve which are calculated by differ-
ing polynomial functions, it would be very surprising if they fitted together to form a smooth
lift for the valve, particularly at the junctions between the elements. In other words, the "jagged
edge" at the calculated boundary between, say, the opening ramp and the main lift, would
provide unacceptably high levels of velocity and acceleration for the valve, if the cam were
43
ground to imitate precisely the calculated valve lift curve. This effect will be demonstrated
numerically in the text and figures below. There is little point in using a "designed" valve lift
curve within an engine simulation, so as to compute the engine performance characteristics, if
that particular valve lift curve cannot be manufactured in practice. To correct this problem, a
simple smoothing routine is easy to formulate and execute at the conclusion of the computa-
tion of a valve lift curve.
Consider an extension of Fig. 1.16, regarding a segment of a valve lift curve, to include the
greater detail shown in Fig. 1.18.
On this segment of a valve lift diagram, there are three points where the lifts are L1, L2,
and L3 at crank angle positions 01, 02, and 03, respectively. Point L2 may possibly be precisely
located on one ofthose "jagged edges" at an inter-polynomial boundary, as mentioned above.
To smooth point 2 so that it fits neatly into the valve lift progression from points 1 to 3, the lift
of point 2 is easily adjusted from L2 to L2a as follows, using similar triangle theory:
L2a = L3- 032(L3 - L) (1.5.46)

03 01
However, as would be conventional in any computation of valve lift, the angular interval
between the three points is normally equal, in which case Eq. 1.5.46 simplifies to:
L2aLI+ L3 (1.5.47)
2
01 02 03
L2a
E L2
E
~LI
CRANK ANGLE, 0Q
Fig. 1.18 Smoothing the computed valve lift profile.
44
To computationally smooth any given valve lift curve, the entire valve lift curve is
indexed, every grouping of three points is sampled, and the lift of each middle point is
adjusted to the mean of the first and last points, according to Eq. 1.5.47.
1.5.6 Numerical Examples of the Computation of Valve Lift andArea
To compute the valve lifts and areas, numbers must be assigned to all ofthe variables seen
in the above equations. Some values refer generically to engine types, i.e., spark-ignition or
compression ignition. Useful data for this purpose are presented below.
Values of the Ramp and Lift Coefficientsfor Spark-Ignition Engines
The numeric values of the coefficients are found from an analysis of measured data from
si engines. The values of the coefficients for the ramps, i.e., kr to kr3, are determined as
0.011172, -0.080336, 0.4686, and 0.6119, respectively. The values of the coefficients for the
main lift, i.e., kmo to km3, are determined as -0.007858, 1.7673, -0.45161, and -0.31158,
respectively. Typical values ofthe lift ratios for the ramps, i.e., both Cur and Cdr as defined in
Eq. 1.5.28, are 0.2 and they are normally equal providing a symmetrical valve lift profile.
Values of the Ramp and Lift Coefficients for Compression-Ignition Engines
The numeric values of the coefficients are found from an analysis ofmeasured data from
ci engines. The values ofthe coefficients for the ramps, i.e., kro to kr3, are found from such an
analysis as -0.00018552, 0.045892, 1.9795, and -1.0256, respectively. The values of the coef-
ficients for the main lift, i.e., kmo to km3, are measured in the same manner as -0.0006083,
1.9236, -0.84122, and -0.082123, respectively. TIypical values of the lift ratios for the ramps,
i.e., both Cur and Cdr as defined in Eq. 1.5.28, are almost double those for si engines, being in
the range 0.4 to 0.5. They are usually equal providing a symmetrical valve lift profile. Diesel
engines normally run slower than gasoline engines so the valves can be opened faster without
incurring excessive velocity or acceleration values.
It will be observed that the zero coefficient, i.e., k,0 or kmo, is not zero in either the si or the
ci case, implying that when Os is zero there is some residual lift present. While this appears
illogical, if one recalls Eq. 1.5.28 and 1.5.45, it will be noted that the valve lift is positively
declared as zero at valve opening and closing, so the potential numeric problem does not arise.
The non-zero value of these coefficients comes about by fitting third-order polynomial curves
to measured valve lift data using mathematical regression techniques.
Comparison ofMeasured and Calculated Valve Lifts
Fig. 1.19 shows the intake valve lift from a four-valve DI diesel engine where the valve
opens at 150 btdc and closes at 380 abdc, giving a total opening duration of 2330 crank angle. It
is a large engine with a 127 mm bore and 135 mm stroke dimensions. The measured intake
valve lift is 11.1 mm. On Fig. 1. 19, the measured data are plotted as a "scatter" graph, i.e., only
the individual points are graphed. The calculated data using the theory from Eqs. 1.5.24-1.5.45
are computed at one-degree intervals and drawn as a line only. The lift ratio for the ramps, up
and down, is 0.45, with this being a diesel engine. The computed valve lift data has been
smoothed ten times sequentially, using the theory from Eqs. 1.5.46 and 1.5.47. It can be seen
that the fit, from calculation to measurement, is very good.
45
11.0
10.0
3 8.0
9.0Z
W
7.0
6.0
W 5.0
< 4.0 *POINTS ARE MEASURED
Z .3.0
2.0 *LINE IS CALCULATED
1.0
0.0* . , . , * , * , *
0 50 100 150 200 250
CRANK ANGLE OF VALVE OPENING,Q
Fig. 1.19 Measured and calculated intake valve liftfor a diesel engine.
The engine runs at 3100 rpm, which is quite a high maximum speed for an engine with
such a long stroke. This engine speed is used in the calculations that also provide the velocity
and acceleration diagrams after the smoothing exercise. They are shown in Fig. 1.20, having
been computed using the theory given in Eqs. 1.5.16-1.5.18. The maximumvelocities are seen
to occur about halfway up the first, and halfway down the final, ramp with a value of 3 m/s.
The maximum acceleration takes place even closer to the commencement of lift, and of valve
closing. The maximum acceleration is just shGrt of 300 g. Nevertheless, small acceleration
ripples can still to be observed at the junctions ofthe ramp and the main lift periods. While this
is possibly unacceptable as a production cam and valve system, any further smoothing needed
for the cam would yield an altered profile involving valve lift changes of the order of microns
(,m). From a simulation standpoint, this would have a quite negligible effect on the valve flow
areas. More detail on this very subject is presented below.
Accuracy of Calculated Valve Flow Areas
Consider an engine that has a bore and stroke of 86 mm, a connecting rod length of
165 mm, runs at 6000 rpm, and has a two-valve hemi-head layout as seen in Fig. 1.2. The
intake valve opens, ivo, at 35 0btdc and closes, ivc, at 75 °abdc. Using the notation of Fig. 1. 15,
it has inner and outer seat diameters, d5s and dos, of 36 and 40 mm, and a valve seat angle, 4, of
45°. The up and down ramp lift ratios, Cur and Cd, are the same at 0.20 and a "ramp up" and a
"ramp down" period of 400 crank angle is used. The dwell period at peak lift is 20 long. The
maximum valve lift, LV, is 13 mm and the duration, v, from the valve timings given above, is
2900. Actually, engines with this cylinder capacity will be featured quite frequently throughout
this entire text, with the data structures for them getting ever more detailed with the passing of
the pages.
46
Chapter I - Introducion to the Four-Stroke Engine
2Z
200j0
7g
ai:
w
0 -Jw
Oo / 100W
w VELOCIT0Y
wU-1 w
> -2
-3 '-100
0 50 100 150 200 250
CRANK ANGLE FROM OPENING, 2
Fig. 1.20 Calculated valve velocity and acceleration diagramsfor a diesel engine.
From a simulation viewpoint, the most important issue is to be able to predict accurately
the valve flow areas. It is seen in Eq. 1.5.5 that there is a simple, albeit inaccurate, way to
obtain the gas flow area presented by a valve as it lifts. An accurate method for this purpose is
presented in Eqs. 1.5.10 and 1.5.12. The valve lift profile for this intake valve data is simulated
by the methods presented above and subjected to ten smoothing sequences. The intake valve
flow areas, Ait, are calculated by both the simplistic treatment and the more accurate method.
The results are shown in Fig. 1.21.
It can be seen that using the simple method of Eq. 1.5.5 will tend to overpredict the valve
flow area by about 10%, i.e., some 130 mm2 on about 1300 mm2 at the peak valve lift point.
Such an error is bad enough. However, when the rest ofthe curve is examined for the error on
flow area, at every point on the lift curve, the simple calculation method is revealed to be
totally inadequate. The error at each point on the lift curve, as a percentage, is calculated by the
following equation:
error =
Ait
100 x simple
-AitItaccnte (1.5.48)
Ataccwatc
This error function is plotted in Fig. 1.22. Here, it can be seen that the error is unacceptable
at the lower values of valve lift where it rises to over 40%, whereas the 10% error at the peak
is really its minimum value. Such inaccuracy, in calculating the valve flow area during a simu-
lation, has serious repercussions for the accuracy of computation ofthe mass flow rate of gas
through it, the filling or emptying of the cylinder contents behind it, and all ofthe thermody-
namic state conditions within the cylinder and in the duct.
47
AREA CALCULATED USING Eqn. 1.5.5

1500
CMJ
E
E
w 1000
0
-J
LL
wi 500
-j
0
0 40 80 120 160 200 240 280
Fig. 1.21 Intake valveflow areas calculated by a simple and an accurate method.
40 ERROR=100*(Asim-Ait)+Ait
U-C
0
-J
30
cc
w
z 20
0
0
cc
LL
w 10
0 40 80 120 160 200 240 280
Fig. 1.22 The simple method ofcalculating valveflow area is inaccurate.
48
The Use of the Smoothing Technique on the Valve Lift Curve

In Fig. 1.18, the technique for smoothing the computed valve lift curve is sketched. Eqs.
1.5.46 and 1.5.48 describe the arithmetic method employed. Fig. 1.23 shows the result of the
smoothing technique for the very intake valve lift data being discussed.
A segment of the computed curve is drawn around the end of the ramp-up period, from 36
to 44 degrees after the valve opens. Without any smoothing of the valve lift curve, the bound-
ary, i.e., the "jagged edge" written about earlier, between the two different polynomials that
detail the ramp up and the main lift up is clearly visible at 40° after the valve opening when the
first ramp concludes. Because the maximum lift is 13 mm, and the ramp lift ratio is 0.2, the end
of ramp lift should be located at 2.6 mm. That such is the case can be seen in Fig. 1.23. On the
same figure is plotted the valve lift curve after the smoothing technique has been applied ten
times. The "jagged edge" is gone. However, this jagged edge amounts to only 0.07 mm at the
410 point, and this difference in lift translates simplistically into a valve flow area correction of
a mere 2.3%. It will be observed that the 2.6 mm lift at the ramp closing point has not been lost
by the smoothing technique.
At the peak lift point, where the main lift up and main lift down segments meet the dwell
period, the "jagged edge" comment seems even more apt. The unsmoothed valve lift curve is
drawn in Fig. 1.24 between 135 and 155 degrees after valve opening.
The peak lift is at 1450 crank angle. The jagged edge is very visible, albeit just 0.03 mm
high. As a result of applying the smoothing technique ten times, plotted on this same figure,
the protrusion has clearly been "filed off," at the expense oflosing 0.03 mm off the peak ofthe
valve lift curve, i.e., a valve flow area loss of about 0.23%.
3.1 LIFT CALCULATED WITH

3.0 1 0SMOOTHING RUNS
2.9
E 2.8 200% RAMP /
E 287 MX F 3m
t~2.6
_J ~NO SMOOTHING
2.5
> 2.4
> 2.3 I
2.2
2.1 2.1-Sf |RAMP
2.0 ! BOUNDARY
1.9- n I w- I . I w J
35 36 37 38 39 40 41 42 43 44
CRANK ANGLE AFTER OPENING, Q
Fig. 1.23 The computational smoothing technique, as applied at a ramp boundary.
49
Design and Simulaton ofFour-Stroke Engines
DWELL
13.00
12.99
12.98 MAIN LIFT UP MAIN UFr DOWN
E 12.97
E
1296w
j 12.95
U 12.94 AFTER 10 SMOOTHINGS
12.93
12.92
12.91 PEAK UFT
12.90 /F| \
12.89
134 139 144 149 154
DEGREE AFTER VALVE OPENING,Q
Fig. 1.24 The computational smoothing technique, as applied at peak lift.
Although the amount of alteration to the valve lift curve is dimensionally minimal, the
effect on the velocity and acceleration curves is very significant. Figs. 1.25 and 1.26 show
graphs of the velocity and acceleration curves for the intake valve operating at 6000 rpm; in
both figures the effect of "no smootiing," and the use of ten applications of the smoothing
technique, is illustrated.
In Fig. 1.25, the effect of the "jagged edges" at the ends of the ramps, and at peak lift, are
clearly seen, as is the result of their removal by the smoothing technique. The velocity incur-
sions that have been deleted are about 1 m/s on amaximum of about 6 m/s, which is about 17%
at the several locations. The high-frequency ripple on the velocity curve is significant, as it
will induce high acceleration rates locally. This is seen to be the case in Fig. 1.26, for the
acceleration incursions run to unacceptably high values of ±3500 g, whereas when the valve
lift curve is smoothed the peak accelerations are reduced to no more than ±700 g.
The smoothing technique is observed to be of real practical significance for the engine
modeller. Using the simulation techniques described throughout this book for the optimization
of the performance characteristics of an engine, the modeller can approach the cam and valve-
train maker with a proposed set of valve lift characteristics that are quite practical as a
manufacturing prospect. At one time, the engine modeller would have been derided by the
mechanical designer for the naivet6 ofhis thoughts on the design ofvalve train geometry. But
not anymore; with these valve lift design techniques those days are history.
50
7 -
6 -
AFTER 10 SMOOTHINGS
CO)
5 -
7. 4 -
/
3 -
2 -
0 1 -
0
-i
w
0 -
-1 .
w -2.
.jl -3.
-4.
-5.
-6.
-7.
0 50 100 150 200 250
CRANK ANGLE AFTER OPENING, Q
Fig. 1.25 The effect of the smoothing technique on valve velocity.
3500 - NO SMOOTHING
0)
z
2500 - / \1
0
1500 -
a:
w 500 -
-1
w
-500 -
w
-1500
/
-i AFTER 10
SMOOTHINGS
-2500
-.^rnn --II
-%-Au%jv v....
I 9 6Iv v 5 8 I 5 5 5 .
I . . . . I .a . . . I . . . .
0 50 100 150 200 250

CRANK ANGLE AFTER OPENING, 0
Fig. 1.26 The effect of the smoothing technique on valve acceleration.
51
Design and SimulWion ofFour-Stroke Engines
The Use ofDiffering Lift Ratios for the Up and Down Ramps
Sec. 1.5.5 describes the possibility of varying the valve lift profile to be asymmetrical.
This is done by setting the lift ratio for the "ramp up" period to be different from that for the
"ramp down" event. To clarify this point, within the data for the intake valve of an si engine
being used as the current arithmetic example, the present data has the ramp lift ratios for the
two ramps, Cur and Cd, as identical at 0.20. These two data values are altered to 0.15 and 0.25,
respectively, and the valve lift profile recalculated. The smoothing technique is executed for
ten times sequentially and the results are plotted in Figs. 1.27-1.29 for valve lift, velocity, and
acceleration.
In Fig. 1.27, the valve lifts for both the symmetrical and the asymmetrical case are plotted
and compared. The lower lift during "ramp up," and the higher lift during the "ramp down,"
are clearly observed and conform to the decreed levels ofthe ramp lift ratios in both cases. The
valve lift is now asymmetrical at 400 and 250° from valve opening, i.e., at the ramp end points,
by a difference of 1.3 mm, yet still retains a visibly smooth profile.
You may well have made the not-illogical mental assumption that a lower value of lift
ratio for the "ramp up" period, i.e., 0.15 instead of 0.20, would correspond to lower levels of
maximum velocity and acceleration for the valve. From Figs. 1.28 and 1.29, this is seen not to
be the case. In Figs. 1.28 and 1.29, although the initial values of velocity and acceleration are
20%RA6MP UP & DOWN

15% UP, 25%RAMP DOWN
13
12
11
10
E 9
Fig 125% RAMP

2
15%/ RAMP
0~
0 40 80 120 160 200 240 280
CRANK ANGLE FROM OPENING,2
Fig. 1.27 Asymmetrical valve lift profiles.
52
7 , 15% RAMP UP
6 / 25% RAMP DOWN
5
c4
51
0 0
-j
w -1
>-2
> -3
<-4
-5
-6 2/°6
20% RMPMUDOWN
RAMP
0 40 80 120 160 200 240 280
CRANK ANGLE FROM OPENING, Q
Fig. 1.28 Velocity diagramsfor asymmetrical valve liftprofiles.
1000 15% RAMP UP

25% RAMP DOWN
0 800
-600
400 0 20%1oRAMP UP
CRANK A 20% RAMP DOWN
Fc
200
Wi 0
> -200
-400
0 40 80 120 160 200 240 280
Fig. 1.29 Acceleration diagramsfor asymmetrical valve lift profiles.
53
indeed lower during the first ramp up, because the valve still must lift to the same maximum
lift of 13 mm, the velocity and acceleration at the beginning of the "main lift up" period now
exceed those when the lift ratio was higher at 0.20. The maximum acceleration on the ramp up
now approaches 1000 g, a rise of some 200 g above that for the symmetrical lift profile. This
could be considered as excessive for a two-valve automobile engine at a peak operating speed
of 6000 rpm. A similar, albeit more obviously intuitive, situation exists at the "ramp down"
period, in the sense that a higher lift ratio ought to provide greater velocities and accelerations.
In short, as within most engineering designs, compromise always must to be sought between
the ideal and the practical.
1.6 Definitions of Thermodynamic Terms Used in Engine Design, Simulation, and
Testing
Throughout this text, the units of any parameter to be inserted into thermodynamic equa-
tions are in strict SI units and all values of pressure and temperature are absolute values. To
ensure a more complete understanding of the tenms in the following discussion, you may find
it helpful at this point to examine Appendix A1. 1, where a precis of some fundamental thermo-
dynamics is presented.
1.6.1 Volumetric Efficiency
In Fig. 1.1(A), the cylinder conducts an intake stroke, in which a mass of fresh air charge,
mav is induced through the intake valve into the cylinder from the atmosphere. This is often
known as the "mass of air supplied," hence the subscript notation. The local atmospheric
conditions of pressure and temperature are conventionally referred to as ambient conditions.
By measuring the local atmospheric pressure and temperature, Pat and Tat, the air density, Pat'
is given by the thermodynamic equation of state, where Ra is the gas constant for air:
Pat =RT (1.6.1)

a at
The volumetric efficiency, T1, of the engine defines the mass of air supplied through the
intake valve during the intake period by comparison with a reference mass, mvref, which is that
mass required to perfectly fill the swept volume under the prevailing atmospheric conditions,
thus:
mwef = PatVsv -lv= Mas (1.6.2)

mvref
1.6.2 Delivery Ratio
In the same cylinder, conducting the same intake stroke, consider the same mass of fresh
charge, mas, being induced through the intake valve into the cylinder from the atmosphere.
The standard reference atmospheric conditions for pressure and temperature, Pdref and Tdref2
54
are defined as being 101325 Pa (1.01325 bar) and 20°C, respectively. The standard reference
air density, Pdref, will be given by the thermodynamic equation of state, where Ra is the gas
constant for air:
Pdref PdRef (1.6.3)
The delivery ratio, DR, of the engine defines the mass of air supplied through the intake
valve during the intake period by comparison with a perfect inhalation into the swept volume
of a standard reference mass of air, m&ef, at the standard reference density, Pdref. The standard
reference mass and delivery ratio can then be expressed as:
mdref = PdrefVsv DR dre

(1.6.4)
The scavenge ratio, SR, of a naturally aspirated engine defines the mass of air supplied
during the intake period by comparison with filling the entire cylinder with a standard refer-
ence mass, msref, that mass which could fill the entire cylinder volume under standard refer-
ence atmospheric conditions:
msref = Pdref(Vsv + Vcv) SR =

(1.6.5)
Msref(165
Should the engine be supercharged or turbocharged, then the new reference mass, mbref,
for the estimation of a scavenge ratio for (what is often referred to as) a blown engine, SRb, is
calculated from the state conditions of pressure and temperature ofthe scavenge air supply, Ps
and T.. This is akin to describing the prevailing atmospheric conditions as being the super-
charger (or turbocharger) line pressure and temperature, and the space being filled by it is the
entire cylinder volume. These definitions devolve to the following equations:
Ps =
Ps
RaTs (1.6.6)
mbref = Ps(Vsv + Vcv) SRb = mbref

ma.6.7)
Some General Comments Regarding Dimensionless Criteria for Air Flow
I prefer to use the definition of scavenge ratio for the naturally aspirated engine in all
circumstances, as given in Eq. 1.6.5, on the grounds that it defines more clearly the amount of
55
air being supplied. For example, ifthe SR value of a naturally aspirated engine is 1.0 and this
engine is then blown to a pressure ratio of 2.0, the SR value, simplistically, would increase to
about 2.0, but the SRb value would remain at unity. I consider that to be both illogical and
confusing, as the greater mass flow rate of air is not visible through the SRb criterion, giving
no expectation of a higher power output to be achieved through supercharging.
The same logic is applied to explain the separate functions required of the criteria of
volumetric efficiency and delivery ratio. The former is used to denote changes in breathing
ability volumetrically, and the latter provides dimensionless information about the mass ofair
ingested. It is important to note that it is air mass, and not volume, that ultimately devolves to
power attained from any engine. Imagine an engine being tested at sea-level at standard atmo-
spheric conditions and then being retested, at the very same engine speed, at an altitude where
the local atmospheric density, Pat, is half that of the standard reference atmospheric value,
Pdref. Let it be assumed that both the volumetric efficiency and the delivery ratio at the sea-
level test are unity for both parameters. At the altitude test, ifthe engine breathed in precisely
the same volume of air, which would not be abnormal, the volumetric efficiency would remain
at unity, but the delivery ratio would be halved (see Eqs. 1.6.1-1.6.4). Thus, the volumetric
efficiency parameter aligns itself with its description, namely it tells us that the engine breath-
ing has been unimpaired with altitude. On the other hand, the delivery ratio criterion informs
us that we can expect to achieve only halfthe sea-level power output ofthe engine. The reason
for this is simple, and will be amplified thoroughly in later discussions. Briefly, because only
half the mass of air is ingested in the test at altitude, we can only burn half the mass of fuel at
the same air-fuel ratio. Hence, if we have available just half the heat energy in the fuel and
bum it at the same combustion efficiency, the end result must approach a halving ofthe power
output.
The SAE standard J604 [1.23] refers to, and defines, delivery ratio. However, the defmi-
tion it provides is precisely that given above for volumetric efficiency. In the literature on four-
stroke engines, delivery ratio and volumetric efficiency are commonly referred to and are
sometimes defined in the fashion of Eqs. 1.6.1-1.6.4, sometimes as in SAE J604, and some-
times are not defined at all within this publication. Thus, one should always be skeptical of
such numbers in a technical paper if the method ofreducing the air flow to the dimensionless
criteria of delivery ratio, volumetric efficiency, scavenge ratio and charging efficiency is not
clearly defined. The definitions given for delivery ratio and volumetric efficiency through
Eqs. 1.6.1-1.6.4 will be adhered to throughout this book.
All of the above theory has been discussed in terms of the ingested air flow compared to
that perfectly inhaled into the cylinder swept volume, as if the engine were only a single-
cylinder unit. However, ifthe engine is a multi-cylinder device, then it is the total air flow into
the entire engine that is referenced to the total swept volume of the engine under consider-
ation, i.e., that defined as Vtsv, in Eq. 1.4.2.
56
1.6.3 Trapping Efficiency

It may be that all of the air supplied to the cylinder, ms, is not trapped within it at intake
valve closure. For example, some may leave through the exhaust valve during the valve over-
lap period. Definitions are found in the literature [1.9, 1.14] for trapping efficiency, TE. The
mass of delivered air that has been trapped is defined as m.s. Trapping efficiency is the cap-
ture ratio of the mass of delivered air that has been trapped by compared to that supplied:
TE = mtas (1.6.8)
mas
1.6.4 Scavenging Efficiency and Charge Purity
The Conditions within a Cylinder
The scavenging efficiency, SE, is defined as the mass ofdelivered air that has been trapped,
mt., compared to the total mass of charge, mt, that is retained at intake valve closure. The
trapped charge is composed of fresh charge trapped, mn,, and the products of combustion
from the previous cycle, mex. These products of combustion have the same composition as the
exhaust gas that leaves the cylinder at release and enters the exhaust system. Semantic confu-
sion tends to arise because in lean combustion not all of the air is consumed during a burn
process and some air remains unbumed from the previous cycle within the "products of com-
bustion," i.e., the "exhaust gas" that departs the exhaust valve when it opens. By the trapping
point when the intake valve closes, the mass of unburned air, ma, retained within the cylinder
must be taken into consideration, as it will be available for, and affect, the next combustion
process. In short, any combustion process takes place between all of the air in the cylinder and
all of the fuel supplied to that cylinder. It is important to define the absolute purity of this
trapped charge. The absolute trapped charge purity, IlabSs is defined as the ratio of air trapped
in the cylinder before combustion, mta, to the total mass of cylinder charge, mtr, where:
=mab-
m
mtas +
Mar0~
mts+mex (1.6.9)
Because the absolute trapped charge purity, fIlabs, within the cylinder is traced during the
course of a simulation by the combination of scavenging efficiency and the separate properties
of the air and the combustion products, i.e., the exhaust gas, it is not featured any further
within this book as a term to be discussed. However, the common tenninology of "purity"
appears very frequently and is required to trace the relative proportions of air and combustion
products, i.e., the exhaust gas, within the cylinder and in the intake and exhaust ducting. You
may well wonder why this is necessary, for the common perception is that the "exhaust gas"
leaving the end of the exhaust tailpipe cannot contain any of the air that ever existed in the
57
intake system, nor can any of the air that exists in the intake ducting ever be contaminated by
"exhaust gas." By the time you have finished this book, such misconceptions will have been
thoroughly dispelled.
Consequently, scavenging efficiency, SE, and the relative charge purity, Hl, have common
definitions for the in-cylinder conditions, both being defined as the relative mass proportions
of the supplied air that has been trapped, mtas, to the totality of the trapped charge:
SE SEHmft
= =m
m=s+ mtw (1.6.10)
mtr Mtas + Mex
In many technical papers and textbooks on two- and four-stroke engines, the word "charge
purity" is often used carelessly by the authors, without the above caveats, which assumes prior
knowledge on the part ofthe reader. Either they assume that all trapped charge is air, or that all
charge supplied is trapped, or that the value of air retained post-combustion, mar, is zero. This
latter assumption may be generally true for most spark-ignition engines and may well be cor-
rect when the combustion process is rich of stoichiometric, but it would not be true for diesel
engines where the air is never totally consumed in the combustion process, and for similar
reasons, it would not be true for a stratified combustion process in a gasoline-fuelled, direct-
injection, spark-ignition (GDI) engine [1.15].
The Conditions within the Engine Ducting
During the valve overlap period, with both valves open, if the intake pressure exceeds the
cylinder and exhaust duct pressures, then air can flow from the intake tract through the cylin-
der into the exhaust pipe. In addition, when the intake valve opens, if the cylinder pressure
exceeds that in the intake system, combustion products, i.e., the "exhaust gas," can enter the
intake duct and mix locally with its air contents. Hence, the totality of what is loosely-termed
"exhaust gas" at some location within an exhaust duct, or "intake air" locally within an intake
system, may well be composed of a mixture of "air" and "exhaust gas", i.e., the combustion
products. During an engine simulation, all gas movements are tracked and the purity of the
charge at any juncture in time and space is one of the properties that is recorded. Consider a
specified zone within the ducting. At some instant in time, it contains a mass of air, maz, and a
mass of exhaust gas, me,,z. The charge purity, HI, is defined as the mass ratio of the air to the
total mass present within the zone, and in a similar manner to that for scavenging efficiency
and relative charge purity within an engine cylinder:
H =
air mass in zone = _Maz
totalmassinzone mat + Mexz(1.6.11)
58
Hence, if the zonal gas present is composed of air only it has a purity of unity, whereas if
it is totally exhaust gas then its purity is zero. Being somewhat knowledgeable about two-
stroke engines where the issue is of even greater importance, I am well aware ofthe semantic
confusion that this topic may cause, so I trust that the thermodynamic debate in Sec. 2.18.9
further clarifies the situation.
1.6.5 Charging Efficiency
Charging efficiency, CE, expresses the ratio of the filling of the cylinder with air, com-
pared to filling the cylinder swept volume perfectly with air at the onset of the compression
stroke. For example, the design objective at the peak power point is to fill the cylinder with the
maximum quantity of air in order to burn a maximum quantity of fuel with that same air.
Hence, charging efficiency, CE, is given by:
CE = mtas (1.6.12)
Mdef
where the reference mass, m&ef, is that defined in Eq. 1.6.4. This may be manipulated to show
that charging efficiency is directly related to delivery ratio and trapping efficiency:
CE=. x -.-
Mas mdref
= TE x DR (1.6.13)
(..3
It should be made quite clear, again, that this is another definition that is not precisely as
defined in SAE J604 [1.23]. In the nomenclature of that SAE standard, the reference mass is
declared to be mvref from Eq. 1.6.2, and not m&ef as used from Eq. 1.6.4. My defense of this
further clash with the SAE standards is the same as that given above for the definition of
delivery ratio, namely that Eq. 1.6.13 reflects a charging efficiency which is basically propor-
tional to the torque produced by the engine, whereas the SAE definition does not necessarily
do that.
1.6.6 Air-to-Fuel Ratio
This subject is discussed in much greater detail in Chapter 4, under the topic of Combus-
tion, but some preliminary data is required to make more informative the Otto and Diesel
cycle theory which is to be found in the later sections of this chapter.
The Gasoline, or Spark-Ignition, Engine
It is important to realize that there are narrow limits of acceptability for the spark-initiated
combustion of air and a fuel such as gasoline. In the case of the gasoline used in the spark-
ignition engine, the ideal fuel is octane, which is the eighth member ofthe family of paraffms
59
Design and Sinulaion ofFour-Stroke Engines
whose general chemical formula is CnH2n+2. Consequently, octane is C8H18, which burns
"perfectly" with air in a balanced chemical relationship, called the stoichiometric equation.
Most students will recall that air is composed, volumetrically and molecularly, of 21 parts
oxygen and 79 parts nitrogen, ignoring the (1%) argon and other trace gases commonly found
in the atmosphere. Hence, the chemical equation for complete combustion becomes:
+ 25 -9 N
79
2C8H18 +2502
L 21J 16C02 18H20 21 2
+ N2 +
(1.6.14)
This produces the information the ideal stoichiometric air-to-fuel ratio, AFR, is such that
for every two molecules ofoctane, twenty-five molecules of air are needed. This information
is normally needed in mass terms. Because the molecular weights of 02, H2, and N2 are
simplistically 32, 2, and 28 respectively, and the atomic weight of carbon C is 12, the air-to-
fuel ratio in mass terms becomes:
(25 x 32) + r25 x 28 x

AFR = 2 1=15.06 (1.6.15)
2(8 x 12 + 18 x 1)
Because the equation is balanced, with the exact amount ofoxygen being supplied to burn
all of the carbon to carbon dioxide and all of the hydrogen to steam, such a buming process
yields the minimum values of carbon monoxide, CO, emission and unburned hydrocarbons,
HC. Mathematically speaking they are zero, and in practice they are also at a minimum level.
Because this equation would also produce the maximum temperature at the conclusion of
combustion, this gives the highest value of emissions of NOX, the various oxides ofnitrogen.
Nitrogen and oxygen combine at high temperatures to give such gases as N20, NO, etc. Such
statements, although based in theory, are almost exactly true in practice, as illustrated by the
discussion in the Appendices to Chapter 4.
As far as combustion limits are concerned, although Chapter 4 will delve into this area
more thoroughly, it may be helpful to point out at this stage that the rich misfire limit of a
spark-initiated gasoline-air combustion probably occurs at an air-fuel ratio of about 9, peak
power output at an air-fuel ratio of about 13, peak thermal efficiency (or minimum specific
fuel consumption) at the stoichiometric air-fuel ratio ofabout 15, and the lean misfire limit at
an air-fuel ratio of about 18. The air-fuel ratios quoted are those in the combustion chamber, at
the onset of combustion of a homogeneous charge, and are referred to as the trapped air-fuel
ratio, AFRt. The air-fuel ratio derived in Eq. 1.6.15 is, more properly, the trapped air-fuel
ratio, AFRt, needed for stoichiometric combustion.
60
To briefly illustrate this point, in the engine shown in Fig. 1.1 it would be quite possible to
scavenge the engine thoroughly with fresh air blown in under pressure and then supply the
appropriate quantity of fuel by direct injection into the cylinder to provide an AFRt of, say, 13.
Due to a generous oversupply of scavenge air, the overall air-fuel ratio, AFRO, could well be in
excess of, say, 20.
The Diesel, or Compression-Ignition, Engine
In the case of the fuel used in the compression-ignition engine, the ideal diesel fuel is
dodecane, the twelfth member of the family of paraffins, C12H26, which has the following
stoichiometric equation:
2C12H26 + 37[02 + N2J

21 24C02 26H20
= + + 37-N22
~~~21 (1.6.16)
(..6
The ideal molecular stoichiometric air-to-fuel ratio for dodecane is such that thirty-seven
molecules ofair are needed for every two molecules of fuel. This translates to an AFR in mass
terns of:
37 x 32 + 37 x 28x 79
AFR = 21 - 14.95 (1.6.17)
2(12 x 12 + 26 x 1)
It will be observed that the stoichiometric AFR for gasoline and diesel fuel are virtually
identical. Unlike gasoline, the rich limit for diesel combustion, in order to avoid an illegal
excess ofcarbon particulates, i.e., a black smoke from the exhaust, is actually well lean of this
stoichiometric AFR. It is at an air-fuel ratio of at least 20, and is probably closer to 23 in most
modem engines designed to meet emissions legislation. Although an AFR of 20 may be the
best available peak power point, the air-fuel ratio for peak thermal efficiency (or minimum
specific fuel consumption) is probably between 23 and 25. The air-fuel ratio at idle (zero net
power output) condition is at about 60. By definition, because the air-fuel ratio is always lean
of stoichiometric, there is air left unused at the conclusion of any diesel combustion process.
Because the fuel is directly injected into the diesel cylinder, and because most diesel
engines are "blown," the trapped air-fuel ratio, AFRt, is nearly always different from the over-
all air-fuel ratio, AFRO; indeed this latter ratio is of very little use in diesel engine design,
development, or simulation.
Equivalence Ratio
In Sec. 1.9.1.5, is a preliminary discussion on the effect ofusing differing air-fuel ratios in
spark-ignition engines, and in Sec. 1.9.2.2 is a similar debate regarding diesel engines. The
concept of equivalence ratio is defined there, which numerically compares any given air-fuel
mixture to its stoichiometric value.
61
1.6.7 Cylinder Trapping Conditions

The point of the foregoing discussion is to make you aware that the net effect of the
cylinder induction process is to fill the cylinder with a mass of air, mt., within a total mass of
charge, mtr, at the trapping point where the intake valve closes. This total mass is highly
dependent on the state conditions at trapping of pressure, ptr, and temperature, Tt1, as the equa-
tion of state shows:
m =RP (1.6.18)
where: Vtr = Vivc + Vcv (1.6.19)
In any given situation, the trapping volume, Vtr, is a constant. This is also true of the gas
constant, Rtr, for gas at the prevailing gas composition at the trapping point. The gas constant
for exhaust gas, Rex, is almost identical to the value for air, Ra. Because the cylinder gas
composition is usually mostly air, the treatment of Rtr as being equal to Ra invokes little error.
For any one trapping process, over a wide variety of breathing behavior, the value of trapping
temperature, Ttr, would rarely change by 5%. Therefore, the value of trapping pressure, ptr, is
the significant variable. As stated earlier, the value of trapping pressure is directly controlled
by the pressure wave dynamics of the intake and exhaust systems, be it a single-cylinder
engine with, or without, tuned intake and exhaust systems, or a multi-cylinder power unit with
a branched exhaust manifold. The methods of design and analysis for such complex systems
are discussed in Chapters 2, 5, and 6. The value of the trapped fuel quantity, mtf, can be
determined from:
m =AFRM (1.6.20)
1.6.8 Heat Released During the Burning Process

The total value of the heat that can be released from the combustion of this quantity offuel
iS QR, and is controlled by combustion efficiency, ilc, and the (lower) calorific value, Cfl, of
the fuel in question:
QR = TnCmtfCfl (1.6.21)
Sec. 1.9.1.6 presents a preliminary discussion of values that can be ascribed to combustion
efficiency for spark-ignition engines, with respect to the air-fuel ratio found during any given
burning process. It will come as no surprise that, in the real world, even the most perfect spark-
initiated combustion process does not provide a combustion efficiency ofunity! On the other
hand, in Sec. 1.9.2 and in Eq. 1.9.9, it can be seen that the combustion efficiency, in the com-
pression ignition of very lean mixtures, does approach unity.
62
1.6.9 The Thermodynamic Cyclefor the Four-Stroke Spark-Ignition Engine

This is often referred to as the Otto cycle, and a full discussion can be found in many
undergraduate textbooks on internal-combustion engines or thermodynamics, such as those
by Obert [1.2], Taylor [1.3], or Heywood [1.25]. Aprecis ofthe thermodynamic theory forthe
ideal Otto cycle is given in Appendix Al.1, Sec. Al.1.8. The result of the calculation of a
theoretical ideal Otto cycle can be observed in Figs. 1.30 and 1.31, by comparison with mea-
sured pressure-volume data from an engine with the same compression ratio. The correlation
between measurement and calculation is not good, which you may well feel is an inauspicious
start to the topic of simulation-in a book that has been declared as being dedicated to this
topic!
In the theoretical case of the ideal Otto cycle-and this is clearly visible on the log(p)-
log(V) plot in Fig. 1.31-the following assumptions are made regarding each stroke of the
cycle:
(a) The compression stroke begins at bdc and is an isentropic, i.e., adiabatic, process.
(b) All heat release (combustion) takes place at constant volume at tdc.
(c) The expansion stroke begins at tdc and is an isentropic, i.e., adiabatic, process.
(d) A heat rejection process (exhaust) occurs at constant volume at bdc.
The compression and expansion processes occur under ideal, or isentropic and adiabatic,
conditions with air as the working fluid, and so those processes are calculated, using the theory
in Appendix A1. 1, Sec. A1. 1.3, as:
pVy = a constant (1.6.22)
160
140
EQUIVALENT IDEAL OTTO CYCLE
120 imep 21.26 bar
ui 100-
a: \ MEASURED (2.OL 14 4v at4800 rpm)
860 imep 14.26 bar
CO
CO
wi 60
a.40
20
0-
0 100 200 300 400 500 600
CYLINDER VOLUME, cm3
Fig. 1.30 Ideal Otto cycle comparison with experimental data.
63
5.5 IDEAL OTTO CYCLE

5.0 MEASURED (2.0L 4v at 4800 rpm)
4.5 n y (ideal O -to)=1.4
4.0
C, 3.5
0
-J2.5
0.5
2.0 /n=1.25
1.5
1.0 'y (ideal Otto)=1.4
0.5
0.0-
3.5 4.5 5.5 6.5
LOGe(V)
Fig. 1.31 Logarithmic plot ofpressure and volume.
The exponent is the ratio of specific heats, y, which has a value for air of 1.4. The theoreti-
cal analysis in Appendix Al.1, Sec. Al.1.8 shows that the thermal efficiency, mnt ofthe cycle is
given by:
,It = CR'Y-' (1.6.23)

where thermal efficiency is defined as:
work produced per cycle (1.6.24)

heat available as input per cycle
In the measured case in Figs. 1.30 and 1.31, the cylinder pressure data is taken from a
2000-cm3 four-cylinder, four-valve, naturally aspirated, spark-ignition, four-stroke engine run-
ning at 4800 rev/min at wide-open throttle. Because the actual 2-liter engine has a compres-
sion value of 10, and from Eq. 1.6.23 the theoretical value of thermal efficiency, mlt is readily
calculated as 0.602, the considerable disparity between fundamental theory and experimenta-
tion becomes apparent, for the measured value is about one-half of that calculated, at around
30%.
Upon closer examination of Figs. 1.30 and 1.31, the theoretical and measured pressure
traces look somewhat similar and the experimental facts do approach some of the presump-
tions of the ideal theory. However, the measured expansion and compression indices are at
64
1.25 and 1.35, respectively, which is rather different from the ideal value of 1.4 for an isen-
tropic process in air. The data can be measured directly from the loge(p)-loge(V) graph as the
value ofa polynomial, or an isentropic, index is the slope ofthe line on a log-log diagram. This
is shown by manipulating Eq. 1.6.22:
loge P = - loge Vy + constant = -y 1oge V + constant (1.6.25)
From Appendix Al.1, Sec. Al.1.6, it is shown for the compression process, where the
polytropic index is 1.35 and is less than the isentropic index of 1.4, that heat is being lost from
the air in the cylinder to the walls. However, during the expansion process, because the poly-
tropic index is 1.25 which is less than y for air at 1.4, there is an implication from the same
theory in Appendix Al.1, Sec. Al.1.6 that heat is being added to the cylinder gas from the
walls. This is not the case. In the real engine, as will be shown in later chapters, the gas in the
cylinder during expansion is not air but combustion products as in Eq. 1.6.14, and the y value
of that very hot burned gas is probably in the region of 1.2. Consequently, heat is still being
lost to the walls during the expansion process.
The theoretical assumption of a constant-volume process for the combustion and exhaust
processes is clearly in error compared to the data on the experimental pressure trace. The peak
cycle pressures, calculated as 160 bar and measured as 70 bar, are demonstrably very different.
Later in this book, but particularly in Chapters 4 and 5, more advanced theoretical analyses
will be seen to approach the measurements ever more closely.
The work on the piston during the cycle is ultimately, and ideally, the work delivered to
the crankshaft by the connecting rod. The word "ideal" in thennodynamic terms means that
the friction or other losses, like leakage past the piston, are not taken into consideration. There-
fore, the ideal work produced per cycle, as in Eq. 1.6.24, is the work carried out on the piston
from the force F, created by the gas pressure p. Work, SW, is always the product of force and
the distance moved by that force, dx. Hence, where A is the piston area:
8W = Fdx = pAdx = pdV (1.6.26)

This is illustrated in Fig. 1.32, where the measured p-V diagram for the 2-liter car engine
is repeated and further information is added.
The element of work, 8W, shown above to be equal to the product, pdV, is seen to be an
elemental area on the pressure-volume diagram. Therefore, the work produced during the
cycle, which is for one crankshaft revolution only from tdc to tdc and includes only the com-
pression and power strokes, is the cyclic integral of these elemental areas over the entirety of
the pressure-volume diagram above the piston in the cylinder. The work, Wi, produced during
the cycle is evaluated over the power mechanical cycle as:
W=
W ;pdV= tdcdc pdV + bdc
td
pdV (1.6.27)
65
75
eo0
60.MEASURED p-V DIAGRAM
dV
CO 3 |NIET CYCLIC WORK= JpdV
CO)
t30
0 100 200 300 400 500

CYLINDER VOLUME, ,.3
Fig. 1. 32 Determination of imep from the cylinder pressure diagram.
Because the value of volume change, SV, during the compression stroke is negative-it is
decreasing with every elemental step-the work put into the compression process is also nega-
tive. The compression work, Wc, is the first half ofthe cyclic integral given above:
tdc
WC =
bdc pdV (1.6.28)
The value ofvolume change, dV, during the power stroke is positive-it is increasing with
every elemental step-so the work produced during the power stroke, i.e., the expansion pro-
cess, is positive. The expansion work on the power stroke, We, is the second half ofthe cyclic
integral given above:
We = dc pdV (1.6.29)
The work produced from the ideal Otto cycle can now be seen as the enclosed area of the
pressure-volume diagram, from the addition of a value of We, which is positive, and Wc,
which is negative. The work produced during the ideal Otto cycle is evaluated only for the
compression and power strokes, because no consideration is given to any work done during
the exhaust or intake strokes, particularly because they are both assumed to have a zero work
content in an ideal engine.
66
Pumping Work on the Exhaust and Intake Strokes

A real engine must inhale air and exhale exhaust gas. The work this takes to accomplish is
called pumping work. The work required to remove the exhaust gas and to induce air must be
taken into account. In Fig. 1.33 is graphed the pumping loop from a real engine, i.e., the
cylinder pressure-volume diagram for the exhaust stroke and the intake stroke.
The pumping work required, Wp, is calculated over the pumping mechanical cycle as:
WpyPd - Wex +Wi

WpgdV=W + in = JbpdV
p + bdc
d pdV (1.6.30)
The pumping cycle takes two strokes, the exhaust and the intake stroke, where the work
required in each stroke, Wex and Win, respectively, makes up the two parts of the cyclic inte-
gral in Eq. 1.6.30. Because the volume change element, dV, is negative on the exhaust stroke
and positive on the intake stroke, the net work required during the pumping cycle is the
enclosed area on the cylinder pressure-volume diagram. When Eq. 1.6.30 is evaluated and Wex
is computed as being a negative number whose absolute value is higher than Wjn, due to the
pressure during the exhaust stroke being higher at all times than during the intake stroke, the
pumping work, Wp, is also a negative number indicating that work is supplied to the engine.
2.75 Vcv
2.50
.0(- 2.25
LL 2.00
cc 1.75
11)
C')
1.50'
1.25
LLI 1.00'
z 0.75- A
J
0.50' E
0.25
0.00
0 100 200 300 400 500 600
Fig. 1.33 Determination ofpmep from the cylinder pressure diagram.
67
In a real engine, the deeper the suction during the intake stroke, or the greater the pressure
bulge during the exhaust stroke, then the larger will be the enclosed area ofthe pumping loop
and the more work will be required of it. To be blunt, the more the spark-ignition engine is
throttled, the more work it takes to suck in a lesser mass of air! To be even more blunt, the
creation of work costs fuel to be bumed and we pay for the fuel.
Pumping work is assumed be zero in the ideal cycle because the cylinder pressures during
the intake stroke and the exhaust stroke are assumed to be both constant and to be equal to each
other. In short, although there is a pumping loop of work that must be done in reality, in the
ideal case it becomes merely a line on the pressure-volume diagram with an enclosed area that
is, by definition, zero.
1.6.10 The Concept ofMean Effective Pressure
As stated above, the enclosed pressure-volume diagram area is the work delivered to, or
provided by, the piston in either the real or the ideal cycle. In Fig. 1.32 the rectangular shaded
area is equal in area to the enclosed cylinder pressure-volume diagram. This rectangle is of
height imep and of length VS,V where imep is known as the indicated mean effective pressure
and Vsv is the swept volume. The indicated mean effective pressure, imep, is evaluated during
the power cycle by:
imep ( wd)
po r
(1.6.31)
Fig. 1.32 shows measured data from an engine where the imep is found to be 14.26 bar.
The shaded area with the height value as imep is drawn to scale on the figure and one can see
the area correspondence with the enclosed power cycle loop.
Similarly, during the pumping work shown in Fig. 1.33, there is a rectangular shaded area
that is equal in area to the enclosed cylinder pressure-volume diagram. This rectangle is of
height pmep and of length Vsv, where pmep is known as the pumping mean effective pressure
and Vsv is, again, the swept volume. The pumping mean effective pressure, pmep, is evaluated
during the pumping cycle by:
pmep =
WP
sv _fpdV)pjjpjng
- sv (1.6.32)
Fig. 1.33 shows measured data from an engine where the pmep is found to be 0.83 bar. On
this figure, the shaded area of height pmep is drawn to scale and the correspondence between
it and the enclosed loop of the pumping cycle can be observed.
68
In the text above, the word "indicated" is introduced as a definition and this may seem a
strange choice ofnomenclature. The word "indicated" stems from the historical fact that pres-
sure transducers for engines used to be called "indicators" and the pressure-volume diagram of
a steam engine, traditionally, was recorded on an "indicator card" [1.2]. In this cylindrical
device with access to the engine cylinder, a pencil is connected to a spring-controlled piston
which moves in correspondence with cylinder pressure. The pencil scribes on a cylindrical
paper card which is made to oscillate by being connected through a cord to the crank as it
rotates. The result is a pressure-volume diagram upon which one applied a planimeter to find
the enclosed area of the work diagram and the imep. As a student, I used one on a low-speed
steam engine quite a few times; the phrase, "an attention-getting experience," comes to mind.
The concept of mean effective pressure is extremely useful in relating one engine devel-
opment to another for, while the units of imep are obviously those of pressure, the value is
almost dimensionless. That remark is sufficiently illogical as to require careful explanation.
The point is, any two engines of equal development or performance status will have identical
values of mean effective pressure, even though they may be of totally dissimilar swept vol-
ume. In other words, if cylinder pressure diagrams are plotted as pressure-volume ratio plots,
and the pressures happened to be identical at equal volume ratios, then the values of imep
attained would also be identical, even for two engines of differing swept volume. You may
glance ahead to Fig. 1.60 to get the point.
1.6.11 Indicated Values of Power and Torque and Fuel Consumption
It will be recalled that the indicated power cycle consists only of the compression and the
power strokes. Power is defined as the rate of doing work. The engine rotation rate is recorded
as revolutions per second, rps, or as revolutions per minute, rpm. Because the four-stroke
engine has a power cycle every two crankshaft revolutions, the power delivered to the piston
crown by the cylinder gas during the power cycle is called the indicated power output, wj,
where, for an engine with n cylinders:
as: W; = net work per cycle x work cycle rate
W, = W x 2
= W x rpm
then: 2 120 (1.6.33)
= nx imep xVs x2 = nx imep xVsv
x
2 120
The indicated torque, Zi, is the turning moment on the crankshaft and is related to indi-
cated power output by the following equation:
Wj =
2icZirps 2nZi* '60
= rpm
rpm = Il'30 (1.6.34)
69
As the engine consumes fuel of a calorific value, Cfl, at the measured (or at a theoretically
calculated) mass flow rate of fff, the indicated thermal efficiency, Tji, of the engine is found
from an extension of Eq. 1.6.24:
power output -1_6_35

- rate of heat input rfCfl (1.6.35)
Of great interest and in common usage in engineering practice is the concept of specific
fuel consumption, i.e., the fuel consumption rate per unit power output. Hence, indicated spe-
cific fuel consumption, isfc, is given by:
f fuel consumption rate - f(6

power output WI
It will be observed from a comparison of Eqs. 1.6.35 and 1.6.36 that thermal efficiency
and specific fuel consumption are inversely related to each other, but without the incorpora-
tion of the calorific value of the fuel. Because most petroleum-based fuels have virtually iden-
tical levels of calorific value, the use of specific fuel consumption as a comparator from one
engine to another, rather than thermal efficiency, is quite logical and is more immediately
helpful to the designer and the developer.
1.7 Laboratory Testing of Engines
1. 7.1 Laboratory Testingfor Power, Torque, Mean Effective Pressure, and Specific Fuel
Consumption
Most of the testing of engines for their performance characteristics takes place under
laboratory conditions. The engine is connected to a power-absorbing device, called a dyna-
mometer, and the performance characteristics of power, torque, fuel consumption rate, and air
consumption rate, at various engine speeds, are recorded. Many texts and papers describe this
process and the Society ofAutomotive Engineers (SAE) provides a test code, J1349, for this
very purpose [1.26]. There is an equivalent test code from the International Standards Organi-
zation in ISO 3046 [1.27].
There is little point in writing at length on the subject ofengine testing and of the correc-
tion of the measured performance characteristics to standard reference pressure and tempera-
ture conditions, because these are covered in the standards and codes already referenced.
A laboratory engine testing facility is diagrammatically presented in Fig. 1.34.
The engine power output is absorbed in the dynamometer, for which the slang word is a
"dyno" or a "brake." The latter word is particularly apt as the original dynamometers were,
literally, friction brakes. The principle of any dynamometer operation is to allow the outer
casing to swing freely. The reaction torque on the casing, which is exactly equal to the net
engine torque, is measured on a lever of length L, from the center line ofthe dynamometer as
70
a force F. This restrains the outside casing from revolving, or the torque and power would not
be absorbed. Consequently, the reaction torque measured is the brake torque, Zb, and is calcu-
lated by:
Zb = F xL (1.7.1)
Therefore, the work output from the engine during an engine revolution is the distance
"travelled" by the force F on a circle ofradius L:
work per revolution = 2nFL = 2iZb (1.7.2)

The measured power output, the brake power, Wb, is the work rate at a rotational speed in
units of revolutions per second, rps, or in units of revolutions per minute, rpm:
Wb = (Work per revolution) x (revolutions / second)

= 27cZbrps = rpm (1.7.3)
To some, this equation may clear up the apparent mystery of the use of the operator X in
the similar theoretical equation, Eq. 1.6.34, when considering the indicated power output, wj.
im & mep
AIR FLOW EXHAUST FLOW

- -o mex
mf
FUEL FLOW
F
Fig. 1.34 Dynamometer test stand recording ofperformance parameters.
71
The brake thermal efficiency, Tib, is then given by the corresponding equation to Eq. 1.6.35:
lb=
power output - Wb 174
rate of heat input mfCfl (1.7.4)
A similar situation holds for brake specific fuel consumption, bsfc, and the corresponding
equation for the indicated values, Eq. 1.6.36:
fuel consumption rate - rf

bsfc =
power output =1 Wb
.
(1.7*55)
However, it is also possible to compute a mean effective pressure corresponding to the
measured power output of an engine with a number of cylinders, n. This is called the brake
mean effective pressure, bmep, and is calculated from a manipulation of Eq. 1.6.33, but in
terms of the measured values:
bmep= Wb = Wb
nx Vsv x ps nx Vsv x (1.7.6)
2 120
It is obvious that the brake power output and the brake mean effective pressure are the
residue of the indicated power output and the indicated mean effective pressure, after the
engine has lost power to internal friction and air pumping effects. These friction and pumping
losses deteriorate the indicated performance characteristics by what is known as the mechani-
cal efficiency ofthe engine, Tlm. Friction and pumping losses are related simply by:
net power = Wb = W; + friction and pumping power (1.7.7)

The mechanical efficiency, Tim, is defined as:
llm = W Wbbmep
= bmep (1.7.8)
Wiimep
This raises the concept of the friction mean effective pressure, finep, which is related
from:
bmep = imep + pmep + fmep (1.7.9)
72
The discering reader may query the veracity of Eq. 1.7.9. However, the pumping mean
effective pressure, pmep, is evaluated from cylinder pressure data by Eqs. 1.6.30 and 1.6.32.
The value of pmep is numerically negative. Historically, the engineer often mentally translates
that into a positive number and attempts to apply it in Eq. 1.7.9 to equate imep, bmep, pmep,
and finep-and fails! Often, the only way to determine the friction content ofthe engine is to
measure imep and pmep with cylinder-mounted pressure transducers, and the bmep on the
dynamometer. These are all measurable parameters on a real engine, but the finep is not nearly
so amenable because much of its value will depend on the very magnitude of the cylinder
pressure loading up the bearings of the connecting rod and the crankshaft
It can be very difficult to segregate the separate contributions of friction and pumping in
measurements taken in a laboratory, except by recording cylinder pressure diagrams in the
manner described above. By measuring friction power using a motoring methodology that
eliminates all pumping or cylinder pressure action at the same time, one negates some of the
friction loss when it is actually pumping and firing. It is very easy to write this out in words,
but it is much more difficult to accomplish this in practice. More information on this topic will
be presented in Chapter 5. One method of measuing mechanical efficiency (reasonably) real-
istically on a dynamometer test-bed for a multi-cylinder engine is to use the Willan's line
method [1.6].
1.7.2 Laboratory TestingforAir Flow Rate and Exhaust Emissions
Air and Fuel Flow
SAE J1088 [1.28] deals with exhaust emission measurements for small utility engines, a
category into which many four-stroke engines fall. Several interesting technical papers have
been published in recent times questioning some of the correction factors used within such test
codes for the prevailing atmospheric conditions. One of these by Sher [1.36] deserves further
study.
The measurement of air flow rate into the engine is often best conducted using meters
designed to conform to a standard, such as the British standard BS 1042 [1.29]. The total air
flow rate into an engine of n cylinders, has, can be reduced to delivery ratio, DR, or scavenge
ratio, SR, values, using Eqs. 1.6.1 to 1.6.5:
DR = 2mw
n x rps x
(1.7.10)
mdref
Upon measurement of the fuel flow rate, ijf, the recording of the overall air-fuel ratio,
AFRO, is relatively straightforward as:
AFRo = mf
(1.7.11)
73
Continuing the discussion begun in Sec. 1.6.5, this overall air-fuel ratio, AFR0, is also the
trapped air-fuel ratio, AFRt, if the engine is charged with a homogeneous supply of air and fuel
that is not lean of the stoichiometric value, i.e., as in a normal carbureted design for a simple
four-stroke engine. If the total fuel supply to the engine is, in any sense, stratified from the
total air supply, AFR will not be equal to AFRt; diesel engines are a classic example of where
this is the case.
Exhaust Emissions
This area is so well covered in the literature that the inclusion of this topic in this book is
merely as an introduction which will serve to assist you to comprehend a huge data base.
Much of the test instrumentation available has been developed for, and specifically oriented
toward, four-stroke cycle engine measurement and analysis.
The legislation on exhaust emissions details the limiting values of all pollutants on a mass
basis at specified power output or load levels. Most exhaust gas analytical devices measure on
a volumetric or molecular basis. It is necessary to convert such numbers from volumetric to
mass values so as to permit the comparison of engines for their effectiveness in reducing
exhaust emissions at equal power levels. Hence, it is necessary to describe here the fundamen-
tal theory required to derive measured, or brake specific, emission values for such pollutants
as carbon monoxide, unburned hydrocarbons, oxides of nitrogen, and carbon dioxide. Ofthese
pollutants, the first is toxic; the second and third are blamed for "smog" formation; the third is
regarded as a major contributor to "acid rain"; and the last is saddled with the blame for a
future "greenhouse effect."
As an example, consider any pollutant gas, G, with molecular weight, Mg, and a volumet-
ric concentration in the exhaust gas of proportion Vcg. Often, such numbers are presented as
parts per million (ppm), or as percentage by volume (%vol). The symbolisms for the unit in
ppm, and the unit in %vol, are Vppmg and V%g, respectively. Their arithmetic connection with
the volumetric ratio is:
volumetric ratio, V = Vppmg 106

= V%g . 100
The average molecular weight of the exhaust gas is Mex. The measured power output is
Wb, and the fuel consumption rate is rjf. The total mass flow rate of exhaust gas is liex, which
is connected by:
mex =
(l+AFRo)mff kg/s
(1 + AFRo)fif kgmol / s (1.7.12)
Mex
74
pollutat gaspollutant
flow rate = (MgVcg) (1 + AFRO)f
~~Mex kg/s (1.7.13)
kgs(7.3
Tlhe brake specific pollutant gas flow rate, bsG, is then given by:
bsGbsG Wb
gc x (I+(Me
AFR)if kg/Ws (1.7.14)
or, bsG =(MgVcg) x ( + AFR0)bsfc kg/Ws (1.7.15)
By quoting an actual example, this last equation is readily transferred into the usual
numeric presentation units for the reference of any exhaust gas. If bsfc is employed in the
conventional units of g/kWh and the pollutant measurement of, say, carbon monoxide, is in %
by volume, the brake specific carbon monoxide emission rate, bsCO, in g/kWh units is given
by:
bsCO = (1+ AFRO)bsfc( J(28- J

go (1.7.16)
where the average molecular weights of exhaust gas and carbon monoxide are assumed sim-
plistically to be 29 and 28, respectively. To ensure total understanding, if in a practical
example of a carbureted engine, the overall AFRO is 13, the bsfc is 290 g/kWh, and the mea-
sured carbon monoxide volumetric concentration in the exhaust is 3%, then the brake specific
carbon monoxide concentration is:
bsCO = 1 + 13) x 290 x 3

x 28) =117.6 g/kWh
The actual mass flow rate ofcarbon monoxide in the exhaust pipe is Iiico where the power
output is Wb in kW units:
imcCo = bsCO x Wb g/h (1.7.17)

It should be pointed out that there are various ways of recording exhaust emissions as
values "equivalent to a reference gas." In the measurement of hydrocarbons, either by an
NDIR (non-dispersive infrared) device or by an FID (flame ionization detector), the readings
75
are quoted as ppm hexane, or ppm methane, respectively. Therefore, the molecular weights for
hexane, C6H14, or methane, CH4, must be inserted in the appropriate equation ifthe pollutant
gas G is to be regarded as equivalent hydrocarbons, HC, to either hexane or methane. Further,
some FID meters use a CH1 85 equivalent [1.28]; in that case, the molecular weight for the HC
equivalent would have to be replaced by 13.85.
The same holds true for nitrogen oxide emissions. It is normally assumed that the mea-
sured value is to be compared to an NO equivalence and so the molecular weight for NO,
which is 30, should be employed. Should the meter used in a particular laboratory be different,
then, as for the HC example quoted above, the correct molecular weight of the reference gas
must be inserted in any application of the above equations.
Trapping Efficiency from Exhaust Gas Analysis
In an engine where the combustion is sufficiently rich, or is balanced as in the stoichio-
metric equation, Eq. 1.6.15, it is a logical assumption that any oxygen in the exhaust gas must
come from air that has been lost through the exhaust valve to the exhaust system during the
overlap valve period. In practice, a stoichiometric air-fuel ratio will have some residual oxy-
gen from combustion sources in the exhaust gas, so such an AFR should not be employed
during measurements of trapping efficiency. In short, the experimental test for trapping effi-
ciency should be conducted with a sufficiently rich mixture during the combustion process as
to ensure that no free oxygen remains within the cylinder after the combustion period. If the
combustion process is deliberately stratified, as in a diesel engine, this test technique cannot
be used. However, on the assumption that the zero oxygen condition for burned gas can be
met-and this is possible because most simple si four-stroke engines are homogeneously
charged-the trapping efficiency, TE, can be calculated from the exhaust gas analysis as
follows:
exhaust gas massflow rate = (I+ AFR )mf kgmls(..8

kg/mols (1.7.18)
Mex
exhaust 02 mass flow rate = Mx %2

100°M (1.7.19)
engine 02 mass inflow rate = 0.2314 mnf AFRO

m 2
kglmol-s (1.7.20)
76
The numerical value of 0.2314 is the mass fraction of oxygen in air and 32 is the molecular
weight of oxygen. The value noted as V%02 is the percent by volume concentration of oxygen
in the exhaust gas.
Trapping efficiency, TE, is defined as:
TE air trapped in cylinder (1.7.21)

air supplied
hence TE 1 air lost to exhaust (1.7.22)

air supplied
from Eqs. 1.7.19-1.7.20
TE = I- + AFR0)V0/M02
(123.14 (1..23
x AFRoMex (1.7.23)
Assuming simplistically that the average molecular weight of exhaust gas is 29, that the
molecular weight of oxygen is 32, and that atmospheric air contains 21% oxygen by volume,
this equation becomes:
TE = 1- ( FR0)V%Q2
2lx AFRO (1.7.24)
The methodology emanates from history in a paper by Watson [1.31] in 1908. Huber
[1.32] basically uses the Watson approach, but provides an analytical solution for trapping
efficiency, particularly for conditions where the combustion process yields some free oxygen.
That it is still an important issue can be seen from more recent publications [1.33].
1.8 Potential Power Output of Four-Stroke Engines
At this stage ofthe book, it will be useful to be able to assess the potential power output of
four-stroke engines. From Eq. 1.7.6, for an engine of n cylinders the power output Wb deliv-
ered at the crankshaft is seen to be:
rb = bmep x n x rm
pv
x2 =bexVtvx120
77
From experimental work on various types of four-stroke engines the potential levels of
attainment of brake mean effective pressure are well known within quite narrow limits. The
literature is full of experimental data for this parameter, and the succeeding chapters of this
book provide further direct information on the subject, often predicted directly by engine
modeling software. Some typical levels of bmep for a brief selection of engine types are given
in Fig. 1.35.
The engines, listed as A-K, can be related to types that are familiar as production devices.
Types A-G are spark-ignition engines running on gasoline; the nomenclature of na (naturally
aspirated), sc (supercharged), and tc (turbo-charged), is quite conventional. The values of mean
piston speed and bmep quoted are at the peak power point in the speed range.
For example, the type A engine could be a lawnmower or a small electric generator of less
than 5 hp. The type B engine would appear in more sophisticated applications in the same field
as engine A, as it is an overhead valve engine, which in the cost-conscious industrial engine
market is a very significant factor. The type C and type D engines would be used for conven-
tional on- or off-road cars or motorcycles. Nowadays, they are also found in outboard motors
in an attempt to meet the oncoming exhaust emissions legislation for marine vehicles. The
type E and type F engines are almost certain to be a car for the sports or luxury car market. The
type G engine is a naturally aspirated racing engine used everywhere from sports car racing to
Formula 1.
Types H-K are compression-ignition engines. The type H engine is a conventional IDI
diesel car engine seen in the European market from about 1975-1995. In more recent times it
has been replaced by turbo-charged IDI and DI engines such as types I and J. The modem
truck engine, as type K, is normally DI and turbo-charged. Many of these engines are also used
as inboard marine units for pleasure and sport fishing boats.
Engine Type bmep Piston Speed Bore/Stroke

bar n/s Ratio
Spark-Ignition
A 2v na sv small industrial 8.0 - 9.0 10 - 12 1.0 - 1.3
B 2v na ohv small industrial 9.0 - 10.0 10 - 12 1.0 - 1.3
C 2v na ohv car/m'cycle 10.0 - 11.0 15 - 18 1.0 - 1.3
D 4v na ohc car/m'cycle 11.0 - 12.0 15 - 18 1.0 - 1.4
E 4v sc ohc car/m'cycle 14.0 - 16.0 15 - 17 1.0 - 1.3
F 4v tc ohc car/m'cycle 14.0 - 16.0 15 - 17 1.0 - 1.3
G 4v na ohc racing car/m'c 12.0 - 15.0 22 - 26 1.2 - 2.2
Compression Ignition
H 2v na IDI car 8.0 - 9.0 11 - 13 0.85 - 1.0
I 3v tc IDI car 12.0 - 14.0 11 - 13 0.85 - 1.0
J 4v tc Dl car 13.0 - 16.0 11 - 13 0.85 - 1.0
K 4v tc Dl truck 14.0 - 17.0 9 - 11 0.80 - 1.0
Fig. 1. 35 Potential performance criteria for some four-stroke engines.
78
Naturally, this table contains only the broadest of classifications and could be expanded
into many subsets, each with a known band of attainment ofbrake mean effective pressure.
Therefore, it is possible to insert these data into Eq. 1.8.1, and for a given engine total
swept volume Vtsv, at a rotation rate rps, determine the power output Wb . It is quite clear that
this might produce some optimistic predictions of engine performance, for example, by
assuming a bmep of 14 bar for a single-cylinder spark-ignition engine of 500 cm3 capacity
running at an improbable speed of 20,000 rpm. However, ifthat engine had ten cylinders, each
of 50 cm3 capacity, it would be mechanically possible to safely rotate it at that speed-as Mr.
Honda proved some thirty years ago in 50cc Grand Prix motorcycle racing! Thus, for any
prediction of power output to be realistic, it becomes necessary to accurately assess the pos-
sible speed of rotation of an engine, based on relevant criteria related to its physical dimen-
sions.
1.8.1 Influence ofPiston Speed on the Engine Rate ofRotation
The maximum speed of rotation of an engine depends on several factors, but the principal
one, as demonstrated by any statistical analysis of known engine behavior, is the mean piston
speed, cp. This is not surprising as a major liting factor in the operation of any engine is the
lubrication of the main cylinder components, the connecting rod, the piston, and the piston
rings. In any given design, the oil film between those components and the cylinder liner will
deteriorate at some particular rubbing velocity, and failure by piston seizure will result. The
mean piston speed, cp, is found from its conventional definition as:
cp = 2 x Lst x rps (1.8.2)
Because one can vary the bore and stroke for any design within a number of cylinders, n,
to produce a given total swept volume, the bore-stroke ratio, CbS, is determined as follows:
Cb db (1.8.3)
The total swept volume of the engine, originally formulated in Eq. 1.4.2, can be written as:
Vtsv n CLt
4 dboLst
=n (1.8.4)
4
Substitution of Eqs. 1.8.2 and 1.8.4 into Eq. 1.8.1 reveals:
=0.333
bmep x cp 0.666 (7nl
Wb = X (CbsX Vts) x - (.8.5>
4 4)
79
This equation is strictly in SI units. Perhaps a more immediately useful equation in famil-
iar working units, where the measured or brake power output, W, is in kW, the bmep is in bar,
and the total swept volume is in cm3 units is:
_ ba
(C bs X
meP0.666
VtSVcm3 1)66
0333
400 (cm3 4J
by removing the constants:
= cp b4335 X (Cbs X Vtsv 31) x (n) kW (1.8.6)
The values for bore-stroke ratio and piston speed, which are typical of the spark-ignition
engines listed as types A-G, are shown in Fig. 1.35. It will be observed that the values of piston
speed, which are quoted at peak horsepower, are normally in a common band from 15-17 m/s
for most spark-ignition engines, and those with values above 22 m/s are for engines for racing
or competition purposes which would have a relatively short lifespan. The values typical of
diesel engines are slightly lower, refleAiavg ijut e ier cylinder components required
to withstand the greater cylinder pressures, but also the reducing combustion efficiency ofthe
Diesel cycle at higher engine speeds and the longer lifespan expected of this type of power
unit. It will be observed that the bore-stroke ratios for petrol engines vary from "square"' at 1.0
to "over-square" at 1.3, with some racing engines going above 2.0. The diesel engine, on the
other hand, has bore-stroke ratios that range in the opposite direction to "under-square,"
reflecting the necessity for a suitable proportioning of the smaller combustion chamber in a
much higher compression ratio power unit.
1.8.2 Influence ofEngine Type on Power Output
With the theory developed in Eqs. 1.8.5 or 1.8.6, it becomes possible by the application of
the bmep, bore-stroke ratio, and piston speed criteria to predict the potential power output of
various types of engines. This type of calculation would be the opening gambit of theoretical
consideration by a designer attempting to meet a required target. Naturally, the statistical
information available would be of a more extensive nature than the broad bands indicated in
Fig. 1.35, and would form what would be termed today as an "expert system." As an example
of the use of such a calculation, four engines are examined by the application of the theory
reflected in Eqs. 1.8.5 or 1.8.6. The results are shown in Fig. 1.36.
The engines are very diverse in character, including, for example, a small lawnmower
engine, a racing motorcycle engine, a truck diesel powerplant, and a Formula 1 racing car
engine. The input and output data for the calculation are declared in Fig. 1.36 and are culled
from those applicable to the type of engine postulated in Fig. 1.35. The target power outputs in
the data table are in kW (horsepower values in brackets). These values are 2.98 kW (4 bhp) for
80
Rotary Racing Trck Fornula 1

Input Data Lawnmower Motorcycle Turbo-Diesel Racina Car
power, kW (bhp) 2.98 (4.0) 123 (165) 224 (300) 522 (7O0)
piston speed, m/s 6.0 25.2 10.0 25.1
bore/stroke ratio 1.0 1.5 0.95 2.1
bmep, bar 9.0 12.8 16.0 12.3
number of cylinders 1 2 6 10
Output Data
bore, mm 53.0 98.4 108.7 92.7
stroke, mm 53.0 65.6 114.4 44.2
swept volume, cm3 110.0 1000.0 6400 2990
engine speed, rpm 3300 11500 2600 17000
Fig. 1.36 Calculation output predictingpotential engine performance.
the lawnmower; 123 kW (165 bhp) for the racing motorcycle engine; 224 kW (300 bhp) for
the truck diesel engine; and 522 kW (7000 bhp) for the Formula 1 car engine. For those readers
who are familiar with lawnrower engines, 1000cc v-twin Superbikes, truck turbo-diesels, or
vlO Formula 1 racing engines, the numbers in the table in Fig. 1.36 have the ring of "bulls-
eyes" scored! An expert system built up this way from measured data on real engines is the
opening gambit in any design process. Designers take note: the more data gathered into an
"expert system," the more logic will appear from its use as a preliminary design tool.
For the initial prediction of the potential power performance of an engine, the most useful
part ofthis method is that considerable pragmatism is injected into the selection of the data for
the physical dimensions and the speed of rotation of the engine.
1.9 The Beginnings of Simulation of the Four-Stroke Engine
To simulate an engine is to mathematically monitor all ofthe unsteady gas flow processes
in, out, and through the entire engine and its ducting, and to track all of the heat transfer and
related thermodynamic effects, such as combustion, in all segments of that ducting and its
cylinders.
In this chapter, the concept of a simple thermodynamic cycle, the ideal Otto cycle, has
already been introduced to trace the events in the closed cycle process constituting the power
portion of the two mechanical cycles that together make up the four-stroke cycle engine. In
Fig. 1.30, it is manifestly clear that the ideal Otto cycle is ineffective in simulating combustion
in a spark-ignition engine, compared to data measured in a real engine. The problem is simply
that an engine cannot conduct combustion at constant volume, i.e., instantaneously at tdc,
because a real buming process takes time, the piston keeps moving, and the cylinder volume
changes. Ifthis latter problem could be remedied by keeping the piston stationary at tdc while
combustion took place and then moving it down on the power stroke when all is burned, the
imep and power would increase by some 50% (see Fig. 1.30). So also would the thermal
efficiency improve as it would require no more fuel per cycle. Needless to add, many have
81
tried to induce stop-go piston motion characteristics into both two-stroke and four-stroke
internal combustion engines. So far none has succeeded, using a plethora of mechanical con-
traptions and linkages, or at least none are known to be in mass production. The tone of these
comments may imply that I think that they never will succeed, not so as I have accepted long
ago that there is no limit to human inventiveness. To the contrary, I would actually encourage
the world's inventors to keep on trying to accomplish this ic-engine equivalent of the "search
for the Holy Grail."
On the assumption that engine simulation should primarily attempt to mimic the real world,
it is essential that a time-related combustion process is modelled. When I was a student, in a
pre-computer age, the university academics taught us about the ideal Otto cycle. Then, they
very sensibly pointed out that further calcu4ation work in this area for a reciprocating engine
was well-nigh useless because the arithmetic required to model a phased combustion process
simply could not be handled in any sensible time frame. Then, the arithmetic time frame was
controlled by a slide rule, log tables, or graph paper. The academics of my day quickly moved
on to the non-reciprocating gas turbine cycle where its simpler arithmetic could be accom-
plished rapidly and a slide-rule could be used to design the engine.
The question is, what does a real combustion process for a spark-ignition engine look
like? Chapter 4 is devoted to this topic. There, it is shown how one can analyze measured
cylinder pressure diagrams and deduce the rate at which heat is released into the cylinder, and
the proportion of fuel burned, as the crank rotates. In Fig. 1.30, and again in Fig. 1.32, is a
measured cylinder pressure diagram from a 2.0 liter automobile engine running at 4800 rpm at
full throttle; the shorthand nomenclature (jargon) for full throttle is often written as wot, i.e.,
wide-open-1hrottle. Using the techniques described fully in Chapter 4, this measured cylinder
diagram is analyzed to determine the rate at which fuel is consumed with time, which is called
the mass fraction burned, Be. The result is shown in Fig. 1.37.
The time dependency ofcombustion is now evident. The spark plug ignited the mixture at
25 °btdc, but nothing transpired until 15 °btdc or 345 °atdc. The interval during which the
flame gets going is called the "delay" period and here it lasts for 100 crank angle. The actual
period of heat release takes place over a burn duration, b, of 50 0crank angle and the mass
fraction burned curve clearly has an exponential profile while it is happening. On Fig. 1.37 are
mentioned two Vibe coefficients to fit an exponential curve, a and m, which are assigned
numbers. Although Chapter 4 could be read at this point, the relevant exponential curve is
repeated here, giving the mass fraction burned, Be, at an angle 0, commencing at the onset of
combustion, i.e., at the end of the delay period:
Be = 1- e (b) (1.9.1)
This information can be programmed into a practical example ofan engine and a compari-
son made ofthe effect it will have on the pressure-volume, and temperature-volume diagram,
by comparison with those predicted by the ideal Otto cycle using the very same geometrical
data. Such a comparison is made in Figs. 1.38 and 1.39.
82
IDEAL OQTO CYC

1.0- CONSTANT m
VOLUME ~I
BURN
I
0.8- wr z
w
z U) P D0)
~z
cc
0.6- co
a:
z z M\
0 0
0.4 - z PHASED BURN detais:-
0 IGNITION 25 9btdc
cc DELAY 10 "crank
U- DELAY
0.2 - BURN DURATION 50 Qcrank
CO)
U)
0.0
335
a
a
-J' 360
i
I
385
VIBE 'a' coefficient 6.0
VIBE 'm' coefficient 2.0
I v
Fig. 1.37 Measured mass fraction burned curvefor a spark-ignition engine.
150 3 3
L.0 125
cn 100
a:
CGO 75
ui
cc
0- 50 PHASED BURN
z
03 25
0
0 100 200 300 400 500 600
Fig. 1.38 P-V curves calculatedfor a practical and an ideal Otto cycle.
83
Design and Simulaion ofFour-Stroke Engines
4000 3
IDEAL OTTO CYCLE
3500
us 3000
D 2500
a: 2000
w
5 1500 4
PHASED BURN
1000
500 2
1
0~
0 100 200 400
300 500 600
Fig. 1.39 T-V curves calculatedfor a practical and an ideal Otto cycle.
Clearly, the use of a "real" combustion curve has a profound influence on the similarity of
the pressure-volume profile to that seen for the real engine in Figs. 1.30 and 1.32. The model-
ing process is obviously getting closer to reality and is now worth pursuing as a design aid.
Why could this not have been done when I was a student? The answer is that it took a computer
to analyze the measured cylinder pressure data to get the burn diagram, calculate the Otto
cycle, and graph the results, all in a sufficiently rapid time frame. If I had tackled this same
arithmetic when I was a student, it would have taken me a month. This practical approach to
combustion simulation is applied to the Otto cycle, where it will be examined in more detail
and its computations compared with those predicted by the ideal Otto cycle. At this early stage
of simulation, we will not be concerned about how the engine gets filled with air, nor gets rid
of the exhaust gas, nor of any scavenging during the valve overlap period. Rather, we will
concentrate on the many fundamental lessons to be learned from basic thermodynamics
applied to the closed system, i.e., the power cycle component ofthe overall four-stroke engine.
1.9.1 Power Cycle Analysis of the Otto Engine
The Test Engine Used as the Spark-Ignition Engine Example
Tlhroughout the rest of this section on the Otto engine, a single cylinder unit of 500 cm3
capacity, running at 4000 rpm, is used to illustrate all ofthe thermodynamic points to be made.
It has a bore of 86 mm, a stroke of 86 mm, and a connecting rod with 150 mm centers. The
gudgeon pin offset is zero. From the bore and stroke values, it is seen to be a "square" engine.
The piston position at any point on the rotation of the crank is found using the theory of Sec.
1.4.4 and the cylinder volume by Eq. 1.4.1. The swept volume is 499.6 cm3, but used in all
thermodynamic equations as 499.6 x 106 m3.
84
A compression ratio of 10 is employed. The clearance volume is obtained from Eq. 1.4.3.
The clearance volume is 55.5 cm3, but used in all thermodynamic equations as 55.5 x 10-6 m3.
The maximum and minimum cylinder volumes, i.e., at bdc and tdc, for this spark-ignition
engine are 555.1 and 55.5 cm3, to be used thermodynamically as 555.1 x 10-6 and 55.5 x
10-6 M3.
The fuel is octane with a declared calorific value of 43.5 MJ/kg. From Eq. 1.6.15, the
stoichiometric air-fuel ratio for octane is 14.95. In all spark-ignition analysis using a phased
burn, the data presented in Fig. 1.37 will be used.
The working fluid within the engine cylinder is considered to be air with the properties
given in Appendix Al. 1, Secs. AI.1.1 and A1. 1.2. However, my own eccentricity in conduct-
ing this type of thermodynamic analysis is to define that, at bdc, the swept volume is regarded
as filled with "fresh" air. This is because the clearance volume is considered to be filled with
exhaust, which is an inert gas, but which also has the properties of air. In short, only the swept
volume at bdc is defined to contain a true "fresh" air, the significance of which theoretical
chicanery will become clear in the section below!
1.9.1.1 Thermodynamic Navigation around the Ideal Otto Cycle
In Figs. 1.38 and 1.39 are seen the four numbered points that mark the events of the cycle.
From Sec. 1.6.9, the four processes that make up the thermodynamic cycle are:
(a) Adiabatic and isentropic compression from points 1-2, where the index of compres-
sion is y
(b) Constant volume heat addition (combustion) from points 2-3
(c) Adiabatic and isentropic expansion from points 3-4, where the index of expansion
is'(
(d) Constant volume heat rejection (exhaust) from points 4-1
The cycle commences at bdc, at point 1. The mass of air in the cylinder is given by Eq.
Al.5 from the declaration that the cylinder contains air at state conditions which are the same
as standard reference atmospheric conditions, i.e., 1.01325 bar and 20°C. Any value of pres-
sure and temperature can be employed, but for this ideal example the reference state condi-
tions are to be used.
Initial Values for Total Mass, Air Mass, Density, and DR
From the state equation, Eq. Al.5, mass of gas in cylinder, mlI:
m-p=V - 101325x555.1x10. = 6.689 x 10-4 kg

RT, 287 x 293
The mass of air trapped in cylinder, mta:
- PlVsv 101325 x 499.6 x 10-6 602x109 k

RT, 287 x 293
85
From Eq. 1.6.1, the reference mass for DR is mdref for which its density is required:
Pat =101325_
Pat =RTt 287 x 293 = 1.205 kg/m3
From Eqs. 1.6.3 and 1.6.4, delivery ratilo is:
DR DR
as mta _ 6.02 x 104 =
1.0
mdref PatVsv 1.205 x 499.6 x 10- =
The delivery ratio, DR, is precisely unity when the swept volume is filled with air at
standard reference temperature and pressure and, by inference, density. This seems logical to
me which explains the "chicanery" alluded to above. However, it is only fair to point out that
many an ideal Otto cycle in many a good textbook does not have starting conditions defined
this way.
Heat Available in Fuel to Be Released at tdc
From Eq. 1.6.20; the mass of fuel trapped, mtf, is found as:
6.2x1-4
mtf = AFR =6.015.06 = 4.0 x 10-5 kg
From Eq. 1.6.2 1, the heat transfer at tdc is equivalent to heat energy in fuel:
Q3 = rlcmtfCfl = 1.0 x 4.0 x 10-5 x 43.5 x 106 = 1740 J
Process 1-2, Adiabatic and Isentropic Compression

From Eqs. A1.26 and A1.50, pressure at end of compression, P2:
P2 =PI(V = pCRY =101325x1014 =101325x25.12
= 25.453 x 105 Pa (ie., 25.453 bar)
86
From Eq. A1 .55, temperature at end of compression, T2:
T2 ) 'CR
=v T2 =293x2.512=736 K
From Eqs. A1.21 and A1.23, the work done during compression is negative:
WI = -miCv(T2 - T1) = -6.689 x 10 x 718 x (736 - 293) = -212.8 J
From Eq. A1.23, the change of internal energy done during compression is positive:
U2 - U1 = mlCv(T2 - T) = 6.689 x 104 x 718 x (736 - 293) 212.8 J
Process 2-3, Constant Volume Combustion

The first law of thermodynamics is applied to the combustion process:
Q3 = U3 - U2 +W = mlCv(T3-T +0
Then solving for T3:
T3=T2+ Q2-=736+ 1740 =4359 K

mICV 6.689x10-4 x718
Using the state equation, as in Eq. Al1.44, solving for p3:
~V2 T3, 4359

P3 =P2 x -xT2=
V2
v3 P2 x T-=25.453 x105 x 736
T2
= 150.75x 105 Pa (ie.,150.75 bar)
Change of intemal energy during the process from 2-3, using "first law" above:
U3 - U2 -Q = 1740 J
87
Process 3-4, Adiabatic and Isentropic Expansion

From Eq. Al.55, temperature at end of expansion, T4, after process 3-4:
T4 V41 ( )1..=-=1 4359

T3 tV3) (C .512
From Eqs. A1.26 and A1.50, pressure at end of expansion, p4:
p4 = P3I p3CRY 10 1 - 6.00 X 105 Pa (which is 6.00 bar)

V3Ji10.
From Eqs. A1.21 and A1.23, the work during expansion is positive:
4= -mCV(T4 - T3) = -6.689 x 10-4 x 718 x (1735 - 4359) = 1260 J

From Eq. Al.23, the change of internal energy done during expansion is negative:
U4 - U3 = mlCv(T4 - T3) = 6.689 x 10-4 x 718 x (1735 - 4359) = -1260 J
Process 4-1, to Complete the Cycle by Constant Volume Heat Rejection

From the first law of thermodynamics:
Q4 =U1 -U4 +W4 =mlCv(T -T4) +0

Change of internal energy during heat rejection from 4-1:
U1 - U4 = Q4 = mlCv(Ti - T4) = 6.689 x 10-4 x 718(293 - 1735) = -693 J
The heat rejected is seen from the above as a negative number:
Ql = -693 J
88
Chapter - Inoduction to the Four-Stroke Engine
Obtaining Net Valuesfor the Cycle

Net work output from the cycle, from Eq. Al.59:
net work = Wn = W12 + W3 = -213 +1260 = 1047 J

Otto cycle thermal efficiency is given by Eq. A1.58:
net work _ W = 1047 02

t heat input Q3 1740
and is also given by Eq. A.1.66:
1-=1- 12.501602
=
and must also be the difference between heat added and rejected:
heat input - heat rejected - 4 1740 - = 1 470.602

heat input 3 1740 1740
The indicated mean effective pressure, imep, can be found using Eq. 1.6.31:
imep = - 1047 _ 20.95 x 105 Pa (i.e., 20.95 bar)

VSV 499.6 x 104
The power output of the engine can be found using Eq. 1.6.33:
W= t xp= 1047 x .120 = 34.9 x 103 W (i.e., 34.9 kW)

-
120
The fuel consumption rate, rf, is found from:
rnf = mass per power cycle x power cycles per second
=MtfX pm 4.0 x 10-5 x 4000 = 1.333 x 10-3 kg / s

120 120
89
From Eq. 1.6.36 the indicated specific fuel consumption can be calculated:
isfc = - 333x 10 3.819 x 10-8 kg/Ws

W- 34.9 x 103fW
However, in more conventional units of kg/kWh, this becomes:
isfc = 3.819 x 10-8 -3.819 x 10-8 x 103 x 3600 kg W s =0._138 kg/kWh

kg _Ws
Ws kW 0.3hk/W
The Fundamentals Check Out Numerically
The assertion in Eq. Al.10 can be seen to be numerically accurate:
S6Q = 6w
2 3 4 1
as: fSQ = J6Q+J6Q+ f6Q+ f8Q =0+1740+0-693 = 1047 J
1 2 3 4
2 3 4 1
and: iSW= J8W+f8W+f8W+f8W=-213+0+1260+0= 1047J
1 2 3 4
and from Eq. Al.1, the cyclic integral of the intemal energy should also be zero:
2 3 4 1
fdU = JdU+ JdU+JdU+fdU =213 +1740-1260-693 = 0 J
1 2 3 4
Figs. A1.3, 1.38, and 1.39 show graphs of pressures and temperatures as a result of a
simulation of the ideal Otto cycle using the above theory and with the same numeric data for
the engine and its initial state conditions at the commencement of the cycle. At the state points
1-4, the very same numeric values of pressure and temperature, as calculated above "by hand,"
are observed on this graph.
Carrying Out the Ideal Otto Cycle Simulation on a Computer
Figs. A1.3, 1.38, and 1.39 show the results of the above calculations for the ideal Otto
cycle, but derived on a computer. The derivation procedure is identical to that formulated
above, but the equations are coded in a calculation language such as Basic or Fortran. If the
90
language has graphics commands, which has been common in Basic on the Macintosh for near
a decade, the graphs such as Figs. 1.38 and 1.39 can be shown directly on the computer screen
with scaled axes, etc., and even made into a "movie" with the piston going to and fro as a
function of either cylinder volume or crank angle. Such a movie is highly educational for
students ofengine design. Engineers always design as an extension ofthe picture in the brain,
and the greater the extent of that picture library, the more inventive
becomes the design process.
For the processes that involve changes of volume, i.e., processes 1-2 and 3-4, which are
conducted above by a hand calculation in one pass from tdc to bdc, the computer will move the
program in short steps of, say, 10 crank angle. At each step, a new piston position and a new
cylinder volume will be found using the theory of Sec. 1.4.4. The altered values of pressure,
temperature, change of work, change of internal energy, and change of heat transfer are com-
puted over the volume increment involved, with the initial state conditions in each case being
the final state conditions at the end of the previous step. At the end of anry one step, the final
state conditions calculated are swapped over to be the initial conditions for the ensuing step.
The work, internal energy, and heat transfer calculated in any one step can each be added to
their own computer store and printed out at the conclusion of the complete cycle. At that point
all of the performce parameters of imep, power, or specific fuiel consumption can be simi-
larly computed. In short, the modem desktop computer can complete the entire simulation of
the thennodynamic cycle, plot its graphics, and run its movie-as fast as I have typed this
sentence.
Graphical Outputffrom a Computer Simulation of the Ideal Otto Cycle
The graphs ofp-V and T-V characteristics shown in Figs. A1.3, 1.38 and 1.39 are already
referred to above. Fig. 1.40 shows the cumulative values of heat, Q, work, W, and intemal
energy, U, as the cycle progresses and with respect to crank angle. The convention in this
chapter is that time starts on tdc at 00 crank angle at the beginning of the intake stroke, so the
compression stroke starts at 1800 and the power stroke commences at the next tdc at 3600
crank angle.
The profile of cumulative work, W, shows it becoming progressively more negative
toward tdc and is -213 J at that point. As the power stroke commences, the cumulative work
becomes zero and then increasingly positive, finishing at +1047 J.
The profile ofcumulative heat, Q, shows it remaining zero until tdc, when 1740 J is added
at that point. During the power stroke this profile is flat at 1740 J. At bdc, the heat rejection
process saps away 693 J, leaving the cumulative heat, Q, value at +1047 J.
The profile of cumulative internal energy, U, shows it becoming progressively positive
toward tdc as the air gets hotter and is +213 J at that point. During the combustion at constant
volume, the 1740 J added from the simulated combustion at constant volume raises it to
1953 J. During'work output on the power stroke, the cylinder charge cools losing 1260 J of
internal energy by bdc, leaving it at +693 J at that point. The final heat rejection process of
-693 J reduces the cumulative internal energy to zero by the end of the cycle.
91
>. 1600 -
1400HETQ
z 1200
w&. 1000
0
-800
600-
-400 -
*
a: 200WOKW
0
200
-200
180 270 360 450 540
CRANKSHAFT ANGLE, deg.
Fig. 1.40 Changes of heat, work, and internal energy in the ideal Otto cycle.
1.9.1.2 Thermodynamic Navigation Around an Otto Cycle with Phased Combustion

The computer simulation is repeated but with the heat addition to the cycle accomplished
using the phased burn represented by the data in Fig. 1.37, otherwise all input data remain
identical to those for the ideal Otto cycle. The only difference in the simulation, as far as the
computer program is concemed, is that ideal adiabatic compression stops at 15 °btdc and
another adiabatic process with intemal heat transfer takes over for the next 500 crank angle.
The intemal heat transfer is the phased combustion process and it spans the end of compres-
sion and the beginning of expansion. The heat input fimction is defined by Eq. 1.9.1 and its
input data are given in Fig. 1.37. Those data specify that ignition is at 25 °btdc, but a delay of
100 is also detailed, so heat input actually begins at 15 °btdc. In terms ofthe previous ideal Otto
cycle diagram, point 2 is now brought back (advanced) to 15 °btdc and point 3 is delayed
(retarded) to 25 °atdc. At all other locations in the cycle, the solution is exactly the same as for
the ideal Otto cycle. During the combustion process between state points 2 and 3, as the simu-
lation proceeds in steps of one degree crank angle, Eq. 1.9.1 is solved for the mass increment
of fuel consumed and Eq. 1.6.21 revisited to obtain the heat input associated with it. The
theory of Appendix Al.l, Sec. A1.1.4 is applied as a solution to what is assumed to be an
adiabatic process in a closed system with both a change of volume and with internal heat
transfer, so the ratio of specific heats remains at y.
Due to the arithmetic complexity involved, I will not bore you, or myself, by setting out
these calculations "by hand," but will simply show, on Figs. 1.38, 1.39, and 1.41, the computer
output of the simulation.
92
Figs. 1.38 and 1.39 show the pressure-volume and temperature-volume graphs when the
phased burn is incorporated into the solution. Here, too, are the same data for the ideal Otto
cycle. At the beginning of this section, it is pointed out that the p-V diagram for the ideal Otto
cycle bears little resemblance to the measured p-V diagram typified by Figs. 1.30 or 1.32, but
the diagram for the phased burn in Fig. 1.38 certainly does. The resemblance does not end
there, for the computed performance parameters for the phased burn Otto cycle are also very
much closer to the reality of measured engine data. In the ideal Otto cycle, the indicated ther-
mal efficiency is 0.602, and the imep and isfc are 20.94 bar and 0.138 kg/kWh, respectively. In
the phased burn Otto cycle the indicated thermal efficiency, imep, and isfc are 0.512, 17.84
bar, and 162 kg/kWh, respectively. To put these number into a practical context, a conven-
tional automobile engine with the same compression ratio and fuel will typically have an
indicated thermal efficiency of 0.45, and imep and isfc values of about 15 bar and
0.190 kg/kWh, respectively. Although the use of a phased burn process has not immediately
produced the same values as a conventional car engine, there has been a major shift toward
realism from those provided by the ideal Otto cycle.
Fig. 1.41 shows the cumulative values of heat, Q, work, W, and intemal energy, U, as the
phased burn cycle progresses and with respect to crank angle. This figure may be compared
with Fig. 1.40 where the equivalent data for the ideal Otto cycle are plotted.
All of the thermodynamic comments above regarding Fig. 1.40 are equally applicable to
Fig. 1.41, except that the cumulative heat transfer, Q, and the cumulative intemal energy change
now have sloped profiles during combustion by comparison with the instantaneous, constant-
volume process of the ideal Otto cycle. Although it is small, the increased negative work
during the compression stroke just before tdc is just discernible.
1.9.1.3 Otto Cycle Data with Respect to lime Rather than Volume
When one employs a pressure transducer in the cylinder of an engine, or indeed anywhere
in an engine, the signal recorded is of pressure with respect to time and not volume as has been
the presentation format thus far in this chapter. Because much of the data in the rest of this
book will compare measured and computed pressure diagrams, you must get used to seeing
them in this fashion. Although the word "time" is mentioned above, rarely is time the x-axis
for a graph; it is normally crank angle at the current rotational speed.
To make the point, the previous pressure and temperature data shown in Figs. 1.38 and
1.39 are redrawn against crank angle, 0, and are presented in Figs. 1.42 and 1.43.
The slope of pressure and temperature rise during combustion in the phased bum cycle is
now very evident compared to the instantaneous process in the ideal Otto cycle. For the phased
burn, the higher pressure before tdc is also very evident at the very end of the compression
stroke, which is caused by heat release commencing before tdc itself. In Figs. 1.42 and 1.43,
and indeed also in Figs. 1.38 and 1.39, the considerable drop in peak pressure and temperature
due to phased combustion should be noted. The peak pressure is almost halved and the peak
temperature has dropped by over 1000IC. No wonder the efficiency and power have deterio-
rated from the ideal; but that is life itself, it is never ideal. The inventor's trick is to attempt to
move closer toward it.
93
OTTo CYCLE
OLSAOCM 0 DlKMIl
' 1400 H-AnObLP DUI¶IN /
1200 -HEAT, Q
W 1000
z
N.. 800
0
I 600 INTERNAL ENERGY, U U
W400
a:
i 200
o 0 } t WORK, W
-200
-400- * * *
180 270 360 450 540
Fig. 1.41 Cumulative heat, work, and internal energy in a "phased burn" Otto cycle.
150 IDEAL OTTO CYCLE
LQU
a 100l
CYCLE
WITH PHASED BURN
CD
CD)
W 50
a
z
180 270 360 450 540

CRANKSHAFT ANGLE, deg
Fig. 1.42 P-6 curves calculatedfor a practical and an ideal Otto cycle.
94
4500 IDEAL OTTO CYCLE

9 4000
Cc 3500 -
< 3000
a:
w
X- 2500
w 2000 -
a:
w 1500 -
Z iooo0 OTTO PHASED

CYCLE BURN
~ 1000WITH
500
0-
180 270 360 450 540
Fig. 1.43 T-6 curves calculatedfor a practical and an ideal Otto cycle.
One of the criteria of practical acceptability of any combustion process is the "rate of
pressure rise" curve during a combustion process. The one for the phased burn is easily
extracted from the computer simulation of the engine cycle. Rate of pressure rise, Apo, is
defined as follows and is nornally obtained at one-degree crank angle intervals:
Ape dp AP
dO -[AOJ1(192 (1.9.2)
It can be seen that it is actually the slope of the p-0 graph at any point, but the incremental
value is normally equally acceptable. The graph for this parameter is conventionally expressed
in bar/deg or atnm/deg units, and is shown in Fig. 1.44.
The maximum value is observed to occur at 5 °atdc at 3.55 bar/deg whereas peak pressure
is 83.4 bar at 17 °atdc. On the same graph is the equivalent curve for a DI diesel engine.
Further discussion on both profiles occurs later in Sec. 1.9.2.2.
1.9.1.4 The Behavior of the Otto Cycle Engine with Differing Compression Ratio
Thermal Efficiency
In Appendix Al.1, Sec. Al.1.8, in Eq. A1.66, it is shown that the higher the compression
ratio, the higher is the cycle efficiency. This equation is plotted as a graph in Fig. 1.45, as it has
already been shown above in the simulation of the ideal Otto cycle that it does yield an indi-
cated thermal efficiency precisely according to this function. The fundamental lesson of this
95
c 4-
cs
L.;
2-
LL
CO
c: OTTo
w
c: 0-
CO)
CO)
cn
w
IL -2- DIESEL
0
LL.
w
I-
a: -4 I....................
310 335 360 385 410
CRANK ANGLE, 2atdc
Fig. 1.44 Rate ofpressure risefor a phased burn in the Otto cycle.
0.65 INDICATED THERMAL EFFICIENCY

z
w
0.60
IL IDEAL OTTO CYCLE
w
-j
0.55
a:
w
I
0.50
w
I
0.45 OTTO CYCLE
WITH PHASED BURN
z
0.40 I I
4 6 8 10 12
COMPRESSION RATIO
Fig. 1.45 Otto cycle thermal efficiency with respect to compression ratio.
96
figure for engine design has been known for many years. As a result, engine R&D, and the
fuels research that accompanies it, has long been orientated to enable a spark-ignition engine
to run at an ever-higher compression ratio. In the 1930s, a typical CR value for the si engine
was 5 or 6. Today, the conventional automobile engine has a compression ratio of about 10.
However, the real engine does not burn according to the ideal Otto cycle. If the engine has
a phased burn, does the indicated thermal efficiency also increase, and in the same proportion
with change of CR as that for the ideal Otto cycle? With a computer-based simulation, we can
easily begin to answer this question. The simulation model with the phased burn is run with
CR data from 5 to 11 and the indicated thermal efficiency output is plotted in Fig. 1.45. It can
be seen that the ideal Otto cycle is more optimistic regarding indicated thermal efficiency, not
just at any one compression ratio, but with increasing compression ratio. It can be seen that Tit
dropped from about 51% to 41% for a compression ratio change from 11 to 4. As seen below,
this means that the power, the torque, and the imep changed by the same proportion.
Cylinder Pressure, Temperature, and Rate ofPressure Rise
Because each of the simulations commences with a DR of unity, the same mass of air is
trapped and the same heat is released, either by a phased burn or in the ideal manner at con-
stant volume. Thus, increasing compression ratio should be reflected in a greater work output,
and higher cylinder pressures and temperatures. Fig. 1.46 shows the peak cylinder pressures
for both the ideal Otto cycle and with a phased burn. It can be seen that, as in the case of the
thermal efficiency, the disparity between them increases very markedly with increasing com-
pression ratio. The profile for the phased burn is very much flatter than that for the ideal cycle,
and at the highest compression ratio the peak pressure in the phased burn is actually less than
half that for the ideal Otto cycle.
Fig. 1.47 shows the peak cycle temperature profiles with compression ratio for both the
ideal Otto cycle and with the phased burn. If anything, the temperature change with compres-
sion ratio in the phased burn has an even flatter profile than that for pressure. With the phased
bum, at a compression ration of 5, the peak cycle mean cylinder temperature is 3070°C, rising
to 3326 °C at a CR of 11, i.e., a rise of 256 'C. In the ideal Otto cycle, the equivalent tempera-
tures at a CR of 5 and 11 are 3776 and 4422 'C, respectively, i.e., a rise of 646°C. These
temperatures are the mean gas temperatures within the cylinder, which is the average of that
gas which is buming and that which is not yet burned. Later, in Chapter 4, the relationship
between this mean temperature and those in the burned and unburned zones will be expanded
much further.
Fig. 1.48 shows the p-V diagrams for all of the compression ratios and with the phased
bum modeling the combustion process. The increasing cylinder pressure with greater com-
pression ratio is easily seen, but the greater enclosed area of the p-V diagram is also very
evident. The higher compression ratio means compression into a smaller volume by tdc, rais-
ing the end of compression pressure and temperature. The end of compression pressure rises
from 9.5 bar when CR is 5; to 29 bar when the compression ratio is 11. That is, compression
pressure virtually triples by doubling the compression ratio.
97
180 - PEAK CYCLE PRESSURE

co
ui 160-
cl) 140-
co) IDEAL OTTO CYCLE
a 120-
C 100-
z
>- 80- 4
< 0 -
OTTO CYCLE
X* WITH PHASED BURN
40- I 1
4 6 8 10 12
COMPRESSION RATIO
Fig. 1.46 Peak cycle cylinder pressure with respect to compression ratio.
0
4500 TEMPERATURE, IDEAL OQTO CYCLE
ui]
a:
0
L
4000
w
w
ui
3500 TEMPERATURE, PHASED BURN
3000~
4 6 8 10 12
COMPRESSION RATIO
Fig. 1.47 Peak cycle cylinder temperature with respect to compression ratio.
98
100
COMPRESSION RATIOS FROM 5 TO 11
EACH WITH SAME PHASED BURN
80 A,
L
Co
CO 60
a..
T- 40 :
w
O
z
>- 20
0 T-
0 100 200 300 400 500 600 700
Fig. 1.48 p- V diagrams for a change ofcompression ratio.
The enclosed area on the p-V diagram is network, thus Figs. 1.49 and 1.50 provide graphical
evidence of the amount of extra work available from the higher CR values.
Fig. 1.49 shows cumulative work with respect to crank angle when a phased burn is used
in the simulation for all CR values from 5 to 11. The extra compression work needed at the
higher CR values can be observed. This work increases from -160 J when the CR is 5 to -225
J when the CR is 11. However, this pays off, for the net work by bdc at the end of the power
cycle has risen to 912 J compared to 707 J, a net gain of 205 J as a result.
In Fig. 1.50, the imep is plotted with respect to compression ratio for both the ideal Otto
cycle and the cycle with the phased burn. The imep in the ideal Otto cycle increases at a higher
rate than that for the phased bum, a trend that complements that already seen for thermal
efficiency in Fig. 1.45. The gain in imep, i.e., both torque and power at the rated speed, is
considerable as they rise by 29% from a CR of 5 to a CR of 11.
Up to now, from the information presented, the designer will be tempted to use a very high
compression ratio-perhaps too high. Thus far, everything points to the highest compression
ratio giving the best answer for thermal efficiency or power output. However, Fig. 1.51 pre-
sents the warning note for those who would incorporate too high a compression ratio in a
spark-ignition engine.
Fig. 1.51 shows the position of peak cylinder pressure, and the rate of pressure rise, for the
phased bum simulations with the compression ratios from 5 to 11. It can be seen that the peak
pressure location gets ever closer to tdc with increasing compression ratio. The rate of pres-
sure rise, Apo, as defined in Eq. 1.9.2, is seen to increase markedly and almost linearly from
99
1000 COMPRESSION RATIOS FROM 5 TO 11

d
11
U-1 EACH-WrTH SAME PHASED . \
a, 800
z
I-
cn 600
5
400
0
w 200
0
5
0 -200
400 -
180 270, 360 450 540
Fig. 1.49 Cumulative cycle workfor a change of compression ratio.
C- 22 -
(is
cn
ax:
CO) 20 -
U)
w
a: 18 -
w
LL
U-
w 16
z
w
LU WITH PHASED BURN
IT
14
4 6 8 10 12
COMPRESSION RATIO
Fig. 1.50 The imepfor the Otto cyclefor a change ofcompression ratio.
100
4 RATE OF PRESSURE RISE 22.8

21 w
19 a
Cul CRN NGEA C-1
4 8C18
20
2 197 .
U3-
O_
17
0
z
w CRANKANGLE AT -16
PEAK PRESSURE c
a: .**.1 15
4 6 8 10 12
COMPRESSION RATIO
Fig.
1...TeEfct 1.P51 Peakpressure
of Usn oyrpcInie location and rate ofpressure
fEpnio n opeso rise.
1.7 to nearly 4.0 bar/deg over the same range of compression ratio. In Chapter 4, on combus-
tion, the rate of pressure rise is cited as a marker for the onset of detonation in a combustion
chamber, a phenomenon that will cause severe mechanical damage to the cylinder compo-
nents. There is a lmit to the amount that compression ratio can be increased in a si engine, and
the rate of pressure rise in the cylnder is one of its principal tracers.
1.9.1.5 The Effect of Using Polytropic Indices of Expansion and Compression
Polytropic Index of Compression
In Appendixv.l1,
A Sec. Al. 1.6, it is shown that if the index of compression is polytropic,
i.e., it 'sn rather than y, where n is less than the ratio of specific heats, y, then heat is lost from
the cylinder during the compression stroke. Naturally, if the index n is greater than y, then the
reverse would be the case, but I cannot think of a single example of such a situation in practice.
Fig. 1.52 shows the effect of running the phased combustion Otto cycle simulation at the
standard data value of compression ratio of 10, but where the index of compression is changed
from its standard value of 1.4, which is the ratio of specific heats, y, for air, to 1. 3, then 1. 2, and
1.1. Adiabatic expansion is retained so the index of expansion is kept at 1.4. Fig. 1.52 shows
the effects on imep and indicated thermal efficiency. Loss of index of compression deterio-
rates both of these values quite significantly, i.e., about 6% over the range tested. It would be
quite normal for a conventional spark-ignition engine to display values for the index of com-
pression as low as 1.25.
101
18.0 0.52
CU z
Li INDICATED ZU
a: 17.8 - THERMAL EFFICIENCY 0
D 1\0.51 E
w
z 17.6 -J
W . 0.50
1-17.4 N
W IMEP
LU 17.2 . 0.414
w z
17.0 . 0.48
1.1 1.2 1.3 1.4 1.5
INDEX OF COMPRESSION
Fig. 1.52 Effect of index of compression on Otto cycle efficiency and work
The manner in which this occurs is best seen in Figs. 1.53 and 1.54, which show cumula-
tive heat and work transfer for the adiabatic case, where y is 1.4, and one of the polytropic
cases, where n is 1.1.
In Fig. 1.53, the loss of heat during the compression stroke is clear, and that carries over to
the expansion stroke so that the cumulative heat is always diminished by that loss. In Fig. 1.54,
the apparent curiosity is that it then takes less work in the polytropic case to compress the
charge to tdc on the compression stroke. Even though the same heat is added during combus-
tion, the cumulative work transfer still ends up short because a lesser peak pressure is attained
during the bum period. The work diagram in Fig. 1.54 looks not unlike the scenario of running
at one of the lower compression ratios, as seen in Fig. 1.49.
Polytropic Index ofExpansion
Appendix Al. 1, Sec. A1. 1.6, shows that ifthe index of expansion is polytropic, i.e., it is n
rather than y, where n is greater than the ratio of specific heats, y, then heat is lost from the
cylinder during the expansion stroke. Ifthe index n is less than y, then the reverse would be the
case. Fig. 1.55 shows the effect of running the phased combustion Otto cycle simulation at the
standard data value of compression ratio of 10, but where the index of expansion is increased
from its standard value of 1.4, which is the ratio of specific heats, y, for air, to 1.425, then 1.45,
1.475, and 1.5. Adiabatic compression is retained so the index of compression is kept at 1.4.
102
Chapter I - Introduction to the Four-Stroke Engin
1600 INDEX OF
COMPRESSION 1.4 -
1400
1200 INDEX OF
cn
w
U- 1000
COMPRESSION 1.1
z 800
if
a:
.-- 600
w 400 INDEX OF
200
0 INDEX OF
COMPRESSION 1.1
180 270 360 450 540
Fig. 1.53 Cumulative heat transferfor a change of index of compression.
900 INDEX OF
800 COMPRESSION 1.4
700
600
w 500 COMPRESSION 1.1
U-
4:
CO) 400
z 300
cn
200 INDEX OF
y 100 COMPRESSION 1.1
0 0
INDEX OF
180 270 360 450 540

Fig. 1.54 Cumulative work transferfor a change of index of compression.
103
. 18 0
0.52
.0 IMEP 0
zL 0.51 z
w
CD 0.50 IL
w 17 wi
0 1 0.49
o= .48 u
WJ 16
U_ INDICATED 0.47 E
LU
o THERMAL
z
.814 EFFICIENCY
1.2.4141.81515 0.46 S
w LU a
z
1 5- * . . .. . -0.45
1.38 1.40 1.42 1.44 1.46 1.48 1.50 1.52
INDEX OF EXPANSION
Fig. 1.55 Effect of index of expansion on Otto cycle efficiency and work.
Fig. 1.55 shows the effects on imep and indicated thermal efficiency. Increase of index of
expansion deteriorates both of these values very significantly, i.e., about 15% over the range
tested. The loss of thennal efficiency, or work output, is very much greater for a 0.1 shift
upward for the index of expansion above the ratio of specific heats, compared to a 0.1 shift
downward for the index of compression.- In short, heat loss on the power stroke has a more
significant effect on efficiency and power than on the compression stroke. In Fig. 1.56, which
displays cumulative work transfer for a ratio of specific heats, y, of 1.4 and a polytropic index
of expansion of 1.5, the loss of net work output is seen to be 97 J. In Fig. 1.54, where the index
of compression shift is 0.3, i.e., from 1.4 to 1.1, the net loss of cumulative work during the
cycle is a mere 40 J by comparison.
Sec. 1.6.9 contains a discussion ofthe numeric value of the index of expansion. This value
is obtained from the analysis of cylinder pressure diagrams and is normally found to be less
than 1.4, which is the ratio of specific heats, y, for air at reference atmospheric conditions. In
that discussion, it is pointed out that (a) the gas in the cylinder of a real engine during expan-
sion is not air, which is how the cylinder gas is being treated in these ideal simulations and (b)
the air, or the post-combustion cylinder gas, is very hot and the ratio of specific heats of any
real gas does decrease with rising temperature. Because the expanding gas in a real engine on
the power stroke probably has a ratio of specific heats, y, of about 1.2, then a measured poly-
tropic index of 1.35 does indicate heat loss from the cylinder during expansion. This estab-
lishes that the fundamental theory of Appendix Al.1, Sec. 1.1.6 is correct in its presumptions
and illustrates the magnitude of deterioration of thermal efficiency, or work output, coming
from numeric differences between the ratio of specific heats and the polytropic indices of
compression or expansion.
104
900 INDEX OF
800 EXPANSION 1.4 -
700
600
Lul 500 \
Z 400 INDEX OF
z 300 EXPANSION 1.5
y 300-
200
o 0 100l
-100
-200
-300, . , . , .
180 270 360 450 540
Fig. 1.56 Cumulative work transferfor a change of index of expansion.
1.9.1.6 The Effect of Using Differing Air-Fuel Ratios in the Simulation

Equivalence Ratio (lambda)
In Sec. 1.6.6, the concept of a stoichiometric air-fuel mixture is established to obtain the
optimum oxidation ofthe carbon and hydrogen elements of a given fuel. Ifthe air-fuel ratio is
rich, then there is insufficient air for all ofthe carbon to be bumed to carbon dioxide, i.e., some
carbon monoxide is produced instead, and so less heat must be liberated per unit mass of fuel
combusted. The amount of heat delivered by a combustion process, for whatever air-fuel ratio,
is already defmed in Eq. 1.6.21. This contains a combustion efficiency term which numeri-
cally controls the effect of buming rich, or lean, mixtures. The finer detail of this topic is the
subject of Chapter 4. Here, an initial study ofthe effect on engine performance of changing the
air-fuel ratio is conducted. To denote the proportion that any given air-fuel mixture is lean, or
rich, of the stoichiometric value, the concept of an equivalence ratio is introduced, and is
frequently referred to in ic enginejargon as "lambda," after the Greek letter that is convention-
ally employed to represent it.
AFR
Equivalence ratio: A'= R (1.9.3)
The air-fuel ratio, AFR, is that found within the cylinder of any given engine at the onset
of combustion, and AFRs is the stoichiometric air-fuel ratio defined, for example, by Eq. 1.6.14
for the burning of octane with air.
105
The combustion efficiency, nc, of a gasoline type fuel such as octane, the combustion
efficiency of which is first mentioned in Eq. 1.6.21, can be expressed in terms of equivalence
ratio from measured data as:
0.75 < X < 1.2 llc = lcmax(-1.6082 + 4.6509k - 2.0764X2) (1.9.4)
where the maximum possible value of the combustion efficiency, lcma, is typically about 0.9
in a spark-ignition engine using a gasoline fuel. The numeric output of Eq. 1.9.4 maximizes at
about 12% lean of stoichiometric. The values calculated by use of Eqs. 1.9.3 and 1.9.4 can be
inserted into Eq. 1.6.21 to determine the total amount of heat to be released at combustion
under any given circumstance.
As with all polynomial equations, these equations are only truly representative over a
limited range of values. In the case of Eq. 1.9.4, the range of effective X values spans normal si
combustion, i.e., from about 0.75 to about 1.20.
Changing Air-Fuel Ratio in the Spark-Ignition Simulation
In the phased burn simulations found thus far in this section, a stoichiometric air-fuel ratio
is used in each case, so Eqs. 1.9.3 and 1.9.4 have also been used inherently within those
several simulation experiments, albeit with the same answer for total heat input for combus-
tion in each example. I apologize for not telling you this earlier, but my excuse is that I have
always found the best form of thermodyilamic education to be one of telling the student a
series of white lies in sequence, as the thermodynamic black truth told all at once is far too
large, and perhaps far too bitter, a pill for anyone to swallow in just one gulp!
Using Eqs. 1.9.3 and 1.9.4 above, and retaining the standard data with a phased burn, the
simulations are repeated for equivalence ratios ranging from 0.75 to 1.1, in steps of 0.05. The
variation ofwork output, graphed as indicated mean effective pressure, imep, is shown in Fig.
1.57. The effects on indicated thermal efficiency, Tt, are shown plotted in Fig. 1.58, and on its
reciprocal relation, indicated specific fuel consumption, isfc, in Fig. 1.59.
Eq. 1.6.21 gives the heat available from combustion, QR. As the air-fuel ratio richens, the
mass of fuel trapped, mt, increases, but the combustion efficiency falls in line with Eq. 1.9.4.
Thus, the heat input reaches a maximum at a particular air-fuel ratio. This occurs at the peak
torque, or imep, point at an equivalence ratio, X, of about 0.8, or some 20% rich of stoichio-
metric. This can be seen in Fig. 1.57 where the peak imep is 19.30 bar when X is 0.8, compared
to 17.84 bar when X is unity. Power at rated speed, torque, or imep are seen to increase by some
8% over that obtained at the stoichiometric mixture. If the mixture is excessively rich, i.e.,
when X is 0.75, some power and torque are lost from the peak torque level. At a 10% lean
mixture, the imep is 16.63 bar, so some 7% torque or power is lost compared to that obtained
at the stoichiometric mixture.
The bad news is that maximum power obtained by fuel enrichment is at the expense of
thermal efficiency and specific fuel consumption. In Fig. 1.58, when X is 0.8 at peak torque,
the thermal efficiency is 0.444, compared to 0.512 at stoichiometric. This is a deterioration of
13% for a gain of 8% on imep, the same trend in fuel consumption as seen in Fig. 1.59. If one
106
20 MEAN EFFECTIVE PRESSURE

m
18 /
W1
a
IMEP
16
C,)
Dh 1 4/
w
CC 12 BMEP
a-
P)w 10 8-
L 6
2 FMEP + PMEP (estimated)

uJ 2-
0.7 0.8 0.9 1.0 1.1 1.2

EQUIVALENCE RATIO, (lambda)
Fig. 1.57 Effect of air-fuel ratio on mean effective pressure.
0.55 INDICATED THERMAL EFFICIENCY
>- 0.50
0
z
O 0.45
LL
w 0BRAKE THERMAL EFFICIENCY
-i
< 0.40
CC
w
I 0.35
0.30 . * * * *
0.7 0.8 0.9 1.0 1.1 1.2
EQUIVALENCE RATIO (lambda)
Fig. 1.58 Effect ofair-fuel ratio on thermal efficiency.
107
-c 280 SPECIFIC FUEL CONSUMPTION
m 260
0
F 240- 2 oF BSFC
D 220
z
o
0 200-
-j
l 180
1 60 -
0-
cn
140'
0.7 0.8 0.9 1.0 1.1 1.2
EQUIVALENCE RATIO, (lambda)
Fig. 1.59 Effect of air-fuel ratio on specificfuel consumption.
calculates the actual fuel consumption rates for the 4000 rpm test point at a stoichiometric
AFR, the power is 34.9 kW and the actual fuel consumption rate is 5.65 kg/h. When X is 0.8
and the AFR is 12.05, the power output is 37.78 kW, the isfc is 0.186 kg/kWh, and the fuel
consumption rate is 7.03 kg/h. Hence, the fuel flow rate to the engine at the maximum power
setting is 24% higher.
1.9.1.7 Including Friction and Pumping Losses in the Simulation
The discussion in Sec. 1.7.1 gives details of the relationship between the indicated work
delivered to the piston compared to that measured on a dynamometer as the brake work.
Eq. 1.7.9 shows that the brake work is the summation of the indicated work and the friction
and pumping losses. Within the simulation described in this section, no knowledge can be
obtained ofthe pumping loss as this simulation ofthe Otto cycle, be it with a phased bum or as
the ideal Otto cycle, is conducted only over the power cycle. Hence, it generates absolutely no
information for the other mechanical cycle ofthe engine where pumping takes place. In other
words, the pumping loop diagram seen in Fig. 1.33 cannot be produced, nor the theory in
Sec. 1.6.9, such as Eq. 1.6.32, indexed.
To demonstrate how brake-related data is found from a simulation, we will assume that
friction and pumping losses are known for our standard engine and the combined total ofthese
losses is 4.46 bar. In other words, and ensuring the numerical clarity required of the use of
Eq. 1.7.9 by using the engineering sign convention for thermodynamics described in Appen-
dix Al . 1, Sec. A1. 1.2, the brake mean effective pressure of the standard engine at the stoichio-
metric AFR, can be found from:
bmep = imep + pmep + fmep (1.9.5)
108
hence: bmep = imep + (pmep + fmep) = 17.84 + (-4.46) = 13.38 bar (1.9.6)
If Eq. 1.9.5 is applied in this fashion to all of the simulations conducted in Sec. 1.9.1.6,
where the AFR is varied widely from rich to lean settings, then an equivalent set of simulation
output becomes available which is related to brake, rather than indicated, data. So, on Figs.
1.57-1.59 are drawn all of the simulation output as brake data to be compared and contrasted
with the indicated data. This output may also be compared with data typical of production si
engines in order to assess how closely this simple simulation is approaching reality. The simu-
lation output of brake-related values is for brake mean effective pressure in Fig. 1.57, for brake
thermal efficiency in Fig. 1.58, and for brake specific fuel consumption in Fig. 1.59.
In Fig. 1.57, the friction and pumping loss at a combined 4.46 bar is seen drawn on the
graph, giving a constant difference of4.46 bar between the imep and bmep lines. When this is
translated into brake thermal efficiency, nb, it can be seen that the effect is no longer a hnear
shift from indicated thermal efficiency, lit. The brake thermal efficiency profile is now very
much flatter through the stoichiometric region, a trend that would be very typical of real si
engines. This same trend is also seen in the bsfc graph in Fig. 1.59.
The peak bmep is now observed to be around 15 bar, a value that is still on the high side for
a real si engine with a delivery ratio of 1.0, but that is not so hugely different from practice.
The best bsfc point of about 220 g/kWh is about 14% too optimistic compared to one of
today's production si automobile engines.
Even Closer to Reality
It will be recalled that these simulation data incorporate input data for the indices of com-
pression and expansion at the ideal value of 1.4, which is equal to the ratio of specific heats for
air. Recalling the discussion in Sec. 1.9.1.4 above, if these indices are each shifted by 0.1, to
1.3 and 1.5, respectively, and the present simulation rerun at a stoichiometric AFR, the bmep
becomes 11.62 bar and the bsfc increases to 247 g/kWh, both ofwhich numbers would be well
in line with current practice.
1.9.2 Power Cycle Analysis of the Diesel Engine
Some Fundamental Differences between the Diesel and the Otto Cycles
Fig. 1.60 shows the measured cylinder traces from an Otto engine and a diesel engine.
They are roughly at the same imep. The diesel engine is a 6.0 liter, six-cylinder, turbocharged,
DI unit, and the Otto engine is a 2.0 liter, four-cylinder, naturally aspirated unit with electronic
fuel injection (EFI). Because one has a cylinder swept volume that is twice the other, and the
diesel engine has a CR of 17 while the Otto engine a CR of 10, the volume scale is drawn as
volume ratio so that the engines may be readily compared. The higher CR ofthe diesel engine
shows up during compression as a volume ratio that is about half that of the Otto engine.
The Otto engine would appear to have combustion at constant volume, although from the
foregoing above we know that not to be the case. On the other hand, the diesel engine has a
much higher end of compression pressure and, more importantly in design terms, temperature.
The diesel burn also seems to be at constant volume, but these first impressions will be dis-
pelled below.
109
140
6.OL 16 4V TC DI DIESEL
sa 120 IMEP=12.43 bar
Li
a 100 2.OL 14 4V NA Si OTTO
(I
CO 80IMEP=14.26 bar
w 80
a- 60
c
W MEASURED CYLINDER PRESSURES
20
o 20S
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
CYLINDER VOLUME RATIO
Fig. 1.60 Measuredp- V diagrams for a diesel and an Otto engine.
There is no denying that the two p-V diagrams are not alike, either in shape or character.
The Otto diagram is short and squat compared to the longer, leaner Diesel diagram.
The Definition of the Ideal Diesel Cycle
The ideal Diesel cycle is presented in some theoretical detail inAppendixAl .1, Sec. A.1.1.0.
It is illustrated there by Figs. A.4 and A1.5, showing pressure-volume characteristics and
temperature-volume characteristics, respectively. There is a common bond in fundamental
thermodynamic thinking between the ideal Diesel and the ideal Otto cycle, the latter being
described in Sec. 1.6.9. The elements of the ideal Diesel cycle are, referring to Figs. A1.4 and
A1.5:
(a) Adiabatic and isentropic compression from state condition 1-2
(b) Constant volume combustion as heat transfer from 2-3
(c) Constant pressure combustion as heat transfer from 3-4
(d) Adiabatic and isentropic compression from state condition 4-5
(e) Constant volume heat rejection to represent exhaust process from 5-1
Thus, there is an extra heat addition process for diesel combustion compared to the ideal
Otto cycle, otherwise the two cycles are identical. The original definition of the ideal Diesel
cycle was that it had only constant pressure combustion whereas that described above is often
called the Dual cycle [1.2]. I define either, or both, as an ideal Diesel cycle.
AppendixAI .1, Sec. Al.1.10 contains all ofthe thermodynamic theory required to resolve
a given design in terms ofits performance characteristics, or its state conditions at points 1-5
throughout the cycle. This theory will be used below to examine a test case for the ideal Diesel
cycle.
110
The Test Engine Used as the Compression-Ignition Engine Example

Throughout this section on the diesel engine, a single cylinder unit of997.5 cm3 capacity,
running at 1800 rpm, is used to illustrate the thermodynamic points to be made. It is, in fact,
one cylinder of a six-cylinder turbocharged DI diesel truck engine of6 liters nominal capacity.
It has a bore of 100 mm, a stroke of 127 mm, and a connecting rod with 217 mm centers. The
gudgeon pin offset is zero. From the bore and stroke values, it is seen to be an "under-square"
engine, the common feature ofwhich has already been commented on in Sec. 1.8.1. The piston
position at any point on the rotation ofthe crank is found using the theory of Sec. 1.4.4 and the
cylinder volume by Eq. 1.4.1. The swept volume is 997.5 cm3, but is used in all thermody-
namic equations as 997.5 x 10-6 m3.
A compression ratio of 17 is used, so the clearance volume is obtained from Eq. 1.4.3. This
clearance volume is 62.3 cm3, but is used in all thermodynamic equations as 62.3 x 10-6 m3.
The maximum and minimum cylinder volumes, i.e., at bdc and tdc, for this diesel engine are
1059.8 and 62.3 cm3, to be used in strict SI units as 1059.8 x 10-6 and 62.3 x 10-6 m3.
The fuel is dodecane with a declared calorific value of 43.5 MJ/kg. From Eq. 1.6.17, the
calculated stoichiometric akr-fuel ratio for dodecane is 14.95. The test engine is to be exam-
ined at a light load point, where the equivalence ratio, X, is 2.95. Consequently, the air-fuel
ratio in the chamber is 44.1. Let it be assumed that 10% of the fuel is to be burned at constant
volume, leaving 90% for combustion at constant pressure.
The working fluid within the engine cylinder is considered to be air with the properties
given in Appendix Al . l, Secs. A1. 1. I and A1.1.2. As stated before in Sec. 1.9.1, I define that
at bdc the swept volume is regarded to be filled with "fresh" air because the clearance volume
is considered to be filled with a gas that is inert, exhaust, but that also has the properties of air.
In short, only the swept volume at bdc is defined to contain a true "fresh" air. Because many of
the data calculated by the simple simulations below are to be compared with measured data
from a firing engine, the state conditions selected at bdc at the onset of the compression stroke
reflect the reality of turbocharging a diesel engine, compared to the standard reference pres-
sure and temperature conditions selected for the naturally aspirated spark-ignition unit. The
state conditions to be used atpoint I in Figs. A1.4 andAl.5 inAppendixAl.1 are 2.33 bar and
70°C. The pressure is known from the measured cylinder pressure data to be 2.33 bar, but the
temperature I have had to estimate based on experience, due to the absence of real data.
1.9.2.1 Thermodynamic Navigation around the Ideal Diesel Cycle
Initial Valuesfor Total Mass, Air Mass, Density, DR, and Heatfrom Fuel at State Point 1
The engine is turbocharged so the state condition 1 is at 2.33 bar and 70°C.
From the state equation, Eq. Al.5, mass of gas in cylinder, ml:
=p1V1 _ 2.33 x 105 x 1059.8 x 10 =25.084 x10 kg

RT1 287 x 343
111
The mass of air trapped in cylinder, mta:
mta=PlVsv = 2.33 x 105 x 997.5 x 10 =23.61x10 kg

RT, 287 x 343
From Eq. 1.6.1, reference mass, mdref, for DR is:
p= Pat - 101325 =1.205 kg/ n3

a RTat 287 x 293
From Eqs. 1.6.3 and 1.6.4, delivery ratio is:
DR- mas mta =

M 23.61 x 10 = 1.96
mdref PatVsv 1.205 x 997.5 x 104
The delivery ratio, DR, is 1.96 and shows the considerable amount of air blown into the
engine from the compressor section of the turbocharger.
From Eq. 1.6.20, the mass of fuel trapped, mtf:
mtf=ta 23.61 x 10 - 5.354x105 kg

From Eq. 1.6.21, total heat transfer around tdc is equivalent to heat energy in fuel:
QR TlCmtfCfl
= = 1.0 x 5.354 x 10 x 43.5 x 106 = 2329 J
As the bum proportion at constant volume, kcv, is decided at 0.1, then from Eq. A1.76:
Q3= kVQR = 0.1 X 2329 = 232.9 J Q = (1-kCV)QR =0.9 x 2329 = 2096.1 J

Process 1-2, Adiabatic and Isentropic Compression
From Eqs. A1.26 and A1.50, pressure at end of compression, P2:
P2 =p4 2-J =plCRY = 2.33x 105 x 17 =2.33x105 x52.8

= 123.02 x 10 Pa (ie., 123.02 bar)
112
From Eq. A1.55, temperature at end of compression, T2:
T__ (VI. ) =(CR)Y- T2 = 343x3.106 = 1065.4 K
From Eqs. A1.21 and A1.23, the work done during compression is negative:
WI = -mCv(T2 - T = -25.084 x 718 x (1065.4 - 343) = -1301.1 J
From Eq. A1.23, the change of intemal energy done during compression is positive:
U2 - Ul = mlCv(T2 - Ti) = 25.084 x 104 x 718 x (1065.4 - 343) = 1301.1 J
Process 2-3, Constant Colume Combustion

The first law of thermodynamics is applied to the constant volume combustion process
where the work output is zero.
Q2 = U3 - U2 +w2 = mlCv(T3-T2 +0
then solving for T3:
T3=T2+ Q2
mICV 1065.4232.9
25.084 x 104 x 718
=1194.7 K
Using the state equation, as in Eq. A1.44, solving for p3:
P2 x -2 T3
P3 X=TPp2X T3 -l=3PO2Xl-
= 123.02 x 105 x 1 94.7
1065.4 = 137.95 x 105 Pa
(i.e., 137.95 bar)
Change of internal energy during the process from 2-3:
U3 - U2 = Q2=2329 J
113
Process 3-4 Constant Pressure Combustion

The first law of thernodynamics is applied to the constant pressure combustion process.
From Eq. A1.77, solving for T4:
T4=T3+ Q__4 =1194.7+ 2096.1 =22.

2026.1 K
m,Cp 25.084 x 10-4 x 1005
From Eq. A1.78, solving for p4:
N = P3 (ie., 137.95 bar)

From Eq. A1.79:
T4 62.3xlO4~x 2026.1 =105.66x I0

V4 = V3 4= m3
T3 1194.7
The change of internal energy during the process from 3-4 is analyzed using the first law
of thermodynamics for the closed system, from Eqs. A1.77 and Al.83:
U4 - U3 = mCv (T4 - T3) = 25.084 x 104 x718x(2026.I-1194.7)= 1497.4 J

By Eq. A1.84 the work output is given by:
4 =
mjR(T4 - T3) = 25.084 x 10-4 x 287 x (2026.1 - 1194.7) = 598.5 J
Process 4-5, Adiabatic and Isentropic Expansion
From Eq. A1.80, temperature at end of expansion, T5, after process 3-4:
__ _ '-.Y ( -0.4
T4 VT T55=62026.10
T226l
1059.8
xl10 =805.6 K
From Eq. A1.81, pressure at end of expansion, p5:
V4 17.95 105 [105.66 xxI11040 = 5.47 x 105' Pa (which is 5.47 bar)

= = 1379 i0V5 t)OS9.8xlO~~~~J
1059.8 =.7Xi5P wihi .7br
114
From Eqs. Al.84, the work done during expansion is positive:
W5= -mCv(T - T4) = -25.084 x 104 x 718 x (805.6 - 2026.1) = 2198 J

From Eq. A1.23, the change of internal energy done during expansion is negative:
U5 - U4 = miCv(T - T4) = 25.084 x 104 x 718 x (805.6 - 2026.1) = -2198 J

Process 5-1, to Complete the Cycle by Constant Volume Heat Rejection
From the first law of thermodynamics, as in Eq. A1.85:
Q, = mICV4Ti - T5) = 25.084 x 104 x 718 x (343 - 805.6) = -833.1 J
Change of internal energy during heat rejection from 4-1:
U1 - U5 = Q = mlCv(T - T5)= -833.1 J

Obtaining Net Values for the Cycle
Net work output from the cycle, from Eqs. A1.70 and A1.71:
Wnet = W2 + W3 + W5 =-1301.1 + 598.5 + 2198 = 1495.4 J
Ideal Diesel cycle thermal efficiency is given by Eq. Al .71:
net work = n - 1495.4

heat input QR 2329
The mean effective pressure, imep, can be found using Eq. 1.6.31 or Eq. A1.86:
me- - 1495.4
imVp 997.5x10@ =14.99x1055 Pa(i.e., 14.99 bar)
VSV 997.5 x 104
The power output of one engine cylinder can be found using Eq. 1.6.33:
rpm 5 1800
W=Wnet X
120
=1495.4 x
120
= 22.43 x 103 W (i.e., 22.43 kW)
115
The fuel consumption rate per cylinder, miif, is found from:
rnf = mass per power cycle x power cycles per second

=x r 1800 = 0.803 x 10-3 kg/ s
12x0=535x(FX120
From Eq. 1.6.36 the specific fuel consumption can be calculated:
iffimf 0.803 x 10-3 8xl8kgW

isfc ==iLOS l =3.58 x1o8 kg/Ws
W 22.43 x 103
However, in more conventional units of kg/kWh, this becomes:
isfc = 3.58 x 10-8 kg 3.58 x 10-8 x 103 x 3600 kg W s

= 0.129 kg/kWh
Ws Ws kW h
Graphical Outputffrom a Computer Simulation of the Ideal Diesel Cycle
Figs. A1.4 and Al.5, inAppendixAl.l, show the results of the above calculations for the
ideal Otto cycle, but derived on a computer. The procedure and comments that are given in
Sec. 1.9.1.1 pertaining to similar computations for the ideal Otto cycle are equally germane
here, so they do not need to be repeated. Because Figs. AA.4 and A1.5 are drawn to scale, the
numbers for pressure, temperature, and volume which are derived above can be seen at the
relevant state positions numbered as 1-5.
Fig. 1.61 shows the cumulative work, heat transfer, and internal energy diagrams for this
ideal Diesel cycle. The 10% jump in heat transfer, and accumulated internal energy, can be
seen at the tdc point where the constant volume heat input takes place. The extra work required
to compress the air in the Diesel cycle, compared to that for the Otto cycle in Fig. 1.40, is
evident. It took just 213 J ofwork to compress the Otto cycle charge at a compression ratio of
10, but it took 1301 J in the diesel engine with the much higher compression ratio of 17.
Although the diesel engine is double the capacity of the gasoline engine, that still translates to
about three times the compression work for an equivalent diesel unit. This explains why a
much more robust starter motor and battery are required to supply the necessary energy to get
a car diesel engine going, as compared to its spark-ignition counterpart.
Ignition ofthe fuel in the diesel engine occurs by heating the injected liquid to vapor and
then heating that vapor to its self-ignition temperature. The end of compression temperature,
calculated above as T2, is 1065.4 K or 792.4°C. A temperature at this level is more than suffi-
cient to carry out the two heating processes described and to initiate ignition. However, under
normal cold-start conditions the initial temperature at bdc would be more like 20°C and a
recalculation of the end of compression temperature gives 910 K, or 637°C; starting is now
more problematic. Under arctic air conditions of -30°C, the equivalent end of compression
116
Chapter 1 - Introducton to the Four-Stroke Engine
3000 IDEAL DIESEL CYCLE

- H Q
-' WITH 10% CV BURN
>- 2500 - WITH 90/ OP BURN
CC 2000-
w 1500 INTERNAL ENERGY, U
0
W 1000
2 500
0i
0 WORK, W
-500
-1000
-t500-
180 270 360 450 540
CRANK ANGLE, Qatdc
Fig. 1. 61 Changes of heat, work, and internal energy in the ideal Diesel cycle.
temperature is 755 K, or 482°C; a conventional starting aid such as a glow plug may need to be
seriously considered.
If the data for the ideal Diesel cycle are adjusted, it becomes possible to approach the
shape of the p-V diagram and the imep output measured for the actual engine. By changing the
input data for both the burn ratio, kcv, and the equivalence ratio, X, to 0.11 and 3.6, respec-
tively, so that both the measured and calculated values of peak cycle pressure and imep coin-
cide, the result is the p-V diagram plotted in Fig. 1.62.
The imep and peak cycle pressures are closely fitted and the calculated diagram is indeed
roughly similar to that measured. However, the calculated values of indicated thermal effi-
ciency and indicated specific fuel consumption, at 0.648 and 0.128 kg/kWh, respectively, are
very wide of the reality mark. The real engine, at an equal load level, would have data for the
same parameters of Tt and isfc at about 0.50 and 0.165 kg/kWh. An error of this order by a
simulation, i.e., over-prediction by some 30%, makes it quite unusable as a design tool.
A second attempt is made by using the classic ideal Diesel cycle where all of the heat
transfer to simulate combustion is applied in a constant pressure process only. The theory for
this is given inAppendixAl.1, Sec. Al. 1.10 using the alternative Eqs. Al .87-Al.89. The input
data for both the burn ratio, kc, and the equivalence ratio, X, are altered to zero and 3.5,
respectively, so that both the measured and calculated value of imep coincides with the mea-
surement. The result is the p-V diagram plotted in Fig. 1.63.
It can be seen that the peak cycle pressure is now poorly correlated, but that will come as
no surprise. The fit of measured and calculated diagram is not as good as in Fig. 1.61, even
though the comparison is being made at equal values of imep. The overall computed values of
117
140
120 CALCULATED, imep=12.41 bar

Lc
C: 100
/ IDEAL DIESEL CYCLE WITH
cn
a:
D i/ 11% BURN AT CONSTANT VOLUME
80
cn 89%/o BURN AT CONSTANT PRESSURE
CO
w 60
z 40
-i
20
0 1 . I ,
I I .
0 200 400 600 800 1000
Fig. 1.62 Matching measured data with a 0.1 kc, burn ratio.
140 / MEASURED, imep=12.43 bar

Cu
(a
120
LLi
a: 100
CO)
CO
w 80 IDEAL DIESEL CYCLE WITH
a: / NO BURN AT CONSTANT VOLUME
a.
a: 60 100%/ BURN AT CONSTANT PRESSURE
w
z
40
-
20
0
0 200 400 600 800 1000
Fig. 1.63 Matching measured data with a constant volume burn only.
118
indicated thermal efficiency and indicated specific fuel consumption, are now 0.636 and 0.130
kg/kWh, respectively, but are still unacceptably optimistic compared to conventional mea-
sured data.
The main problem is that the heat input profile that gives the measured data in Figs. 1.62
and 1.63 is quite different from that envisaged for either version ofthe ideal Diesel cycle, i.e.,
as either a constant pressure process only, or with some ofthat total heat input also released in
a constant volume process.
1.9.2.2 Thermodynamic Navigation around the Diesel Cycle with Phased Combustion
The theoretical approach typified by Eq. 1.9.1, which is explained in great detail in
Chapter 4, provides the mass fraction burned diagram in Fig. 1.37 from measured cylinder
pressure-volume data of a 2.0 liter, four-cylinder, spark-ignition Otto engine. The measured
pressure-volume data, shown in Fig. 1.60, are taken from the 6.0 liter turbocharged diesel
engine, and is analyzed using the same techniques. The mass fraction burned diagram
obtained for the diesel engine is shown in Fig. 1.64. It will be recalled that the engine load
level is quite modest, being at 12.43 bar imep, which is obtained at an equivalence ratio of
about 3, whereas maximum usable power will be at ax level ofabout 1.65. The word "usable,"
here means that the engine performs without having an illegal excess of black smoke, i.e.,
carbon particulates, in the exhaust gas.
1.0
MEASURED OTTO
CX0X9 ,
u 00 M8 O DIESEL MEASURED
z 0.
D 0.7 IDEAL DIESEL
m
Z 0.6 06/10%CV BURN
0
O 0.5 VIBE BURN PROFILES
< 0.4 engine Diesel Otto
U 03 start, 9btdc 14 25
ol)< 0.3elay9
0.2 bum,Q
6
70
10
50
Vibe'a' 7.15 6.0
0.1 Vibe'm' 0.69 2.0
0.0 . . I . .
I v I
345 360 375 390 405 420

CRANK ANGLE, 2atdc
Fig. 1. 64 Mass fraction burnedfor actual engines andfor an ideal Diesel cycle.
119
To make comparisons more meaningful, on this same Fig. 1.64 is repeated the measured
mass fraction burned curve for the spark-ignition engine as seen in Fig. 1.37. Also on Fig. 1.64
is drawn the mass fraction burned data which emanates from the very same ideal Diesel cycle
analysis that produced Figs. A1.4 and A1.5, in Appendix A1. 1, and Fig. 1.61. A summary of
the numeric data for the Vibe functions of the measured Diesel and Otto burn curves is pre-
sented on Fig. 1.64; the symbolism for these numbers can be seen in Eq. 1.9.1 and from the
discussion associated with this equation.
To further emphasize the point, the mass fraction burned curve is transformed into a heat
release rate profile. Before you reach Sec. 4.2, which discusses the fundamental theory regard-
ing the relationship between heat release rate, QRe, and mass fraction burned, B, the following
simple equations will suffice to make the introduction. Ifthe amount of heat added by combus-
tion is aQR0, in any given crank angle interval dO, at a crank angle 0, from the beginning of
the bum process, then the heat release rate is defined as:
QRO - 8QRe
dO J/deg (1.9.7)
The mass fraction burned, B(, at any given angle 0, from initiation ofcombustion in a total
burn period b is given by:
IX° QR
=
Xa8QR0
Bo B=
QR (1.9.8)
X8QRO
0
So, mass fraction burned is the cumulative integral under the heat release rate curve to any
angular period 0 compared to that for the entire (crankshaft) angular period, b. For the total
burn period it is the total amount released, QR, as defined by Eq. 1.6.21. Using this theory, the
mass fraction bumed curves, as measured for these Otto and diesel engines, are redrawn as the
heat release rate diagrams graphed in Fig. 1.65; they are both rendered as specific diagrams to
make comparison easier.
The first observation that can be made is that the measured mass fraction burned profiles
for the diesel and Otto engine appear to be somewhat similar. However, an examination ofthe
numeric data for their Vibe functions show that, while the shape may be somewhat similar, the
only data value of any real similarity is the 50% value which occurs at the same location at
some 9 °atdc. Otherwise, compared to that for the diesel engine, the Otto burn starts some 80
earlier, lasts for 200 longer, and has Vibe exponents that are significantly different. The more
rapid input of heat around the tdc location for the diesel engine is more noticeable on the heat
release diagram in Fig. 1.65, as is the long tail ofslow end burning which is equally obvious in
Fig. 1.64. What is very clear is that the mass fraction burned diagram for the ideal diesel cycle
120
w -
0.9 4'
cc
wu 0.8
CD * HEAT RELEASE RATES
0.7
I / FROM MEASURED DATA
0.6
a:
W0.4
W o DIESEL
~0.2/345 36O35T9T45O2
CRAN ANLE QUOd
0.0
345 360 375 390 405 420
CRANK ANGLE, 9atdc
Fig. 1.65 Measured heat release rate curvesfrom Otto and diesel engines.
has virtually nothing in common with that from measured data. The entire heat input rate is
much too rapid, so it is perhaps not too su'rprising that its incorporation into a simulation gives
only a modest correspondence with the measured p-V diagram and the overall performance
characteristics.
The next step is to replay the measured mass fraction burned diagram into the simulation
and find out, to what degree, it has improved the accuracy of this model of a diesel engine.
Calculation ofthe Diesel Cycle with Phased Combustion
The simulation is programmed to receive a phased burn as a Vibe function, and alternate
changes of indices of compression and expansion, otherwise it is identical with the ideal cycle
computation. The computer simulation is conceptually identical to that discussed above for
the Otto cycle with a phased burn. The Vibe coefficient function data, as seen in Fig. 1.64, is
used as input data. The index of compression is also an input data value and that found in the
measured Diesel p-V diagram, i.e., 1.33, is inserted into the simulation. The index of expan-
sion, as with the Otto cycle experiments earlier in Sec. 1.9.1.4, is raised above the ideal ratio of
specific heats so that heat loss is simulated in the expansion process on the power stroke; a
modest value of 1.45 is used as a data value. For completeness, the measured index of expan-
sion is 1.35, but of course is not used as a data value in this simulation for all of the same
reasons expounded in the discussions in Sec. 1.9.1.4 and Appendix Al.1, Sec. Al.1.6. Other-
wise, all of the numeric data for the engine, given at the very beginning of this section (Sec.
1.9.2), are put into the input computer data files.
The results of the simulation, showing both measured and computed pressure-crank angle
diagrams, are shown in Fig. 1.66.
121
140
MEASURED
.6 120 &
La CALCULATED
a:ioo
CD)
W, 80
a-
a: 60
Z 40-
20
0 20
180 270 360 450 540

CRANK ANGLE, gatdc
Fig. 1.66 Cylinder pressure-measured and calculated using a phased burn.
Because the simulation closely mimics the measured pressure-crank angle diagram, it
becomes difficult to determine the level of accuracy attained. Hence, Fig. 1.67 is created,
which shows the pressure difference between the two graphs in Fig. 1.66 In Fig. 1.67 it is
observed that the majority of the error is around the tdc point at the end of compression and the
beginning of combustion.
To illustrate further the conclusions of Sec. 1.9.1.4 regarding heat loss due to indices of
compression and expansion not being identical to the ideal ratio of specific heats, the cumula-
tive work, heat, and internal energy diagrams are drawn in Fig. 1.68.
The cumulative heat, Q, graphed in Fig. 1.68 illustrates the heat loss during compression.
This goes negative until the heat input during the simulated combustion, then decreases again
with further heat loss during expansion, until the cycle concludes with the final heat rejection
"exhaust" process. The effect is considerable and can be seen by comparing Fig. 1.68 with Fig.
1.61, i.e., for the ideal Diesel cycle with the same heat input at tdc. The maximum cumulative
heat input rises to just over 2000 J compared to 2329 J when the compression and expansion
processes are adiabatic. The final maximum value ofcumulative heat for the phased bum case
is 1836 J, so 493 J, or 21%, is lost from the cylinder during the compression and expansion
processes. Due to heat loss from the cylinder, the internal energy of the cylinder gas never
attains the level it did through the adiabatic processes in the ideal cycle. The maximum inter-
nal energy attained in Fig. 1.61 is 3000 J and in Fig. 1.68 it is 2142 J. Such is the heat transfer
realism that must be incorporated into any simulation which has the necessary pretensions of
accuracy to become a useful design tool.
Considering the close correlation of measured and calculated pressure diagrams in Fig.
1.66, unsurprisingly the predicted value of indicated mean effective pressure, imep, and that
measured are coincidental at 12.43 bar. The predicted value of indicated specific fiuel con-
sumption is 0.155 kg/kWh, i.e., some 20% worse than for the ideal Diesel cycle computations
122
L..
7
6 PRESSURE DIFFERENCE
0
a: MEASURED-CALCULATED
5
w 4
w
cDn: 3
CO)
CO 2
w
cc
a. 1
z 0
0
-1
-2
-3 -
180 270 360 450 540
CRANK ANGLE, 2atdc
Fig. 1.67 Difference between measured and calculated cylinder pressure.
2500 1DIESEL CYCLE WITH PHASED BURN

2000
INTERNAL ENERGY, U\
lc: 1500
w
z
w 1000
L-
o
500
I 0
~i
0 -500
-1000
-1500 -
180 270 360 450 540
CRANK ANGLE, Qatdc
Fig. 1. 68 Cumulative work, heat, and internal energy using a phased burn.
123
as seen from Figs. 1.62 and 1.63. "Worse" in this context means closer to reality, as the indi-
cated thermal efficiency has dropped from an over-optimistic 0.648 to a more-realistic 0.533.
Further comment on this point continues below.
Phased Combustion with Varying Fuel Injection Quantity
The diesel engine controls load solely by altering the injected fuel quantity between that
which provides a zero work output, i.e., the idle condition, and a full load condition, i.e., that
which is normally limited by the amount of black smoke permitted by law to appear in the
exhaust gas. To fulfill these two criteria, the fuelling will range between an equivalence ratio,
X, of about 4.5 to about 1.5. The equivalence ratio is seen to be always lean of stoichiometric.
Over this fueling range, the combustion efficiency, ilc, as defined in Eq. 1.6.21, for a diesel
fuel such as dodecane, can be culled from measured data and expressed as a function in terms
of equivalence ratio as:
1.33 < X < 2.33 nc = 0.35332 + 0.56797X - 0.12472x2 (1.9.9)

X 2 2.33 11c 1.0
This expression is not an absolute "law" and would have a somewhat, but not hugely,
different characteristic from one particular DI diesel engine to another, and between DI and
IDI engines in general.
If the fueling level is set so as to attain the maximum possible power at an equivalence
ratio of about 1.4, the maximum cylinder pressure encountered may well exceed that value of
cylinder pressure, and also temperature by inference, which is regarded as safe from the stand-
point of durability of the cylinder components. In practice, the timing of the onset of fuel
injection, and the rate at which it is then injected, are controlled so that a predetermined maxi-
mum cylinder pressure limit is not exceeded. Modem technological advances in electronics
make this form of fueling control more accurate and precise. To illustrate these points, the
equivalence ratio, X, is changed in a series of computational experiments from 3.0, to 2.75,
2.5, 2.25, 2.0, 1.85, and 1.65. None of the other data values pertaining here is changed, with
the exception that Eq. 1.9.9 is employed to find the combustion efficiency, Tic, at any given
fueling level, and the timing point for the start of fuel injection is retarded to ensure that the
same maximum cylinder pressure is reached at every fuelling condition. The start of fuel
injection is set to 14, 13, 12, 11, 10, 9, and 8 'btdc, respectively, for each of the equivalence
ratios shown. The shape and delay of the mass fraction burned diagram is retained, as is the
amount of ignition delay, but the previous statement changes the point of fuel injection from
346 to 347,348,349,350,351, and 352 °atdc, respectively. Although this is not precisely what
would happen in practice, the effect within this simple cycle simulation is sufficiently close to
reality to illustrate the effect on an engine.
Fig. 1.69 shows that the result of controlling the injection timing is indeed a relatively
constant maximum cylinder pressure with variable fueling. In this figure, all of the simulated
pressure-crank angle diagrams are shown over the burn period. The effect of injection delay
on the shape of the cylinder pressure profile with enrichment is quite clear. At the richest
124
trVl IKInCO DO=C:QI II2=

140 k = 3.0 wLTLINUr- rnc-OOUnr
FOR X=-1.65-3.0
co 120
c 101065
50
C')
WJ 80
a.
cc
wu 60
a
z
>.40
Fi1.9Clndrpesredarmsfo hae un6ih5rinflln
00
350 360 370 380 390 400 410 420 430
CRANK ANGLE, gatdc
Fig. 1.69 Cylinder pressure diagrams from a phased burn with varying fueling.
setting, where X is 1.65, the cylinder pressure only begins to rise at about the tdc point. The
rate of pressure rise recorded is about 3.8 bar/deg, which makes it at least as high as the full
load point for the Otto engine seen in Fig. 1.44 and, in effect, at least as noisy. That it is
actually a noisier combustion process is best seen in Fig. 1.44, where the rates of pressure rise
are plotted at full load for both the Otto and the diesel engine. For the diesel engine, the next
derivative of this parameter (in bar/deg/deg units) has a value that is at least double that for the
Otto engine. These rapid vibrations, transmitted through the metal of the cylinder head and
block, provides the "rattle" that is so typical of Diesel combustion.
The mean cylinder gas temperature diagrams resulting from the simulation are plotted in
Fig. 1.70. It is observed that, although the maximum cylinder pressure has been kept relatively
constant, the maximum cycle temperature has not. The difference in maximum cycle tempera-
ture from the lowest to the highest loading is some 500°C.
The effect of fueling on the overall performance parameters for the indicated values of
mean effective pressure, imep, and specific fuel consumption, isfc, and thermal efficiency, 9t,
are shown in Figs. 1.71, 1.72, and 1.73, respectively. The imep profile in Fig. 1.71 shows a
steady increase with fueling enrichment and the torque peak has not yet been reached by an
equivalence ratio of 1.65; it would do so when X approached 1.3, but the black smoke level in
the exhaust would make the engine environmentally unacceptable. This trend is predicted by
the graph of indicated thermal efficiency in Fig. 1.73 which shows that it begins to drop off as
the rich limit approaches, due to the application ofEq. 1.9.9 within the simulation. The recip-
rocal picture of this comment is given in Fig. 1.72, where the specific fuel consumption starts
to rise as the same rich limit approaches. At the leaner air-fuel ratio settings, i.e., when X is
greater than 2.0, the isfc and Tit curves are relatively flat.
125
KA,AKI tWI Wn=D

fl T=KAP=DATI 10=
i) 1800 - FOR X1.65-3.0
Li 1600
cc X=1.65
1400 /
W 1200
(-
z 1000
a: 800
z 600
>'400/
Z 200
180 270 360 450 540

CRANK ANGLE, "atdc
Fig. 1. 70 Cylinder temperature diagramsfrom a phased burn with varying fueling.
22 MEAN EFFECTIVE PRESSURE

m 20
T- 18
CO) 16 IE
w
cc14
wL 12
> 10
8
w
z 4 A A A A A
Wi 2
0
1.50 1.75 2.00 2.25 2.50 2.75 3.00
Fig. 1. 71 The effects offueling on mean effective pressure.
126
250 - iPECIFIC FUEL CONSUMPTION

i
0
225 BSFC
cn
z 200
C)
0
0
w
U-
0~ 175
cn
LL
ISFC\
UJ &a 0
150 - 1
1.50 1.75 2.00 2.25 2.50 2.75 3.00
Fig. 1. 72 The effects offiueling on specificfuel consumption.
0.55 -
0.50 -
U INDICATED THERMAL EFFICIENCY
z
w
U- 0.45 -
w
-J
0.40
w BRAKE THERMAL EFFICIENCY
I
0.35 -
II 0
O.':on
A..
JV I v
I * I * I * I *
i
1.50 1.75 2.00 2.25 2.50 2.775 3.00

Fig. 1. 73 The effects offueling on thermal efficiency.
127
Including Friction and Pumping Losses in the Diesel Simulation

In a similar fashion to the simulation of the Otto engine described in Sec. 1.9.1.6, it is
possible to assign to the above diesel engine a friction and pumping loss and so translate all
simulation data from indicated parameters to brake-related parameters.
It will be assumed that friction and pumping losses are known for our standard diesel
engine and the combined total of these losses is 4.46 bar. For the sake of a clear comparison,
this number is retained to be exactly the same as that used for the Otto engine in Sec. 1.9.1.6,
even though the numeric value for a diesel engine is generally somewhat higher than for an
Otto engine. If the pumping and friction loss is applied in the manner of Eq. 1.9.6, as a com-
bined fmep and pmep to all of the Diesel simulations above where the equivalence ratio is
varied from rich to lean, then another set of simulation output appears which is related to
brake, rather than indicated, data. So, on Figs. 1.71-1.73 is drawn the further simulation output
of brake-related data. The graphs of brake-related values are for brake mean effective pressure
in Fig. 1.71, for brake specific fiuel consumption in Fig. 1.72, and for brake thermal efficiency
in Fig. 1.73.
In Fig. 1.71, the consistent 4.46-bar difference appears on the plot, with the maximum
usable bmep at 17.0 bar, a number that is quite consistent with current turbocharged DI prac-
tice. In Fig. 1.72, a different trend appears for bsfc, compared to the trend commented on
above for isfc. Now, the best brake specific fuel consumption is closer to the highest usable
load point and deteriorates by some 20% toward the lightest load point. In Fig. 1.73, the recip-
rocal comment applies, i.e., the best brake thermal efficiency is near the highest load point and
drops as the load level decreases by reduced fuelling. The value of best point specific fuel
consumption is 200 g/kWh, and its associated thermal efficiency is 41%, two values that are
absolutely in line with those obtained by current, turbocharged, DI diesel engines used in
trucks.
1.10 The End of the Beginning of Simulation of the Four-Stroke Engine
You have now been introduced to the concept of the thermodynamic simulation of an
engine through the use ofthe ideal Otto and Diesel cycles. These ideal cycles are shown to be
rather inaccurate for the purpose of design, although the fundamental lessons in thermody-
namics are invaluable in pointing the way forward to enhancing that accuracy.
The first major step in this regard is the introduction of a combustion process that results
from the analysis of measured cylinder pressure data and that relates the happenings in the
combustion chamber to the buming of a real fuel with respect to time. When this alone is
introduced into the simulation of the ideal cycle, be it Otto or Diesel, the cylinder pressure
diagrams and the predicted overall perfornance parameters begin to more closely coincide
with measured data.
The second major step forward is to introduce heat transfer loss into the simulation of the
ideal cycle which, of course, makes it no longer an ideal cycle but moves it ever closer to
reality. This heat transfer loss is effected by using indices of expansion and compression that
differ from the ratio of specific heats and so negates the assumption that those processes are
ideal, i.e., they are adiabatic and isentropic.
The third, and final, step forward is to assign friction and pumping losses to the engine
cycle and transform the simulation output from indicated performance parameters to brake
128
related data. That step gives engine performance characteristics for bmep and brake specific
fuel consumption that are very similar to that measured for the Otto and diesel engines used as
computation examples.
So If the Cycle Simulation Is Now That Good, What Is the Rest of this Book About?
Firstly, all simulation above is carried out by assuming the trapping state conditions of
pressure and temperature at the beginning of compression. In short, the other half ofthe four-
stroke cycle has been ignored, thus far. A simulation is required that will predict those very
state conditions at the trapping point on the compression stroke, which is really at intake valve
closure and not at bdc. The simulation must be extended to include all of the gas flow through-
out all of the ducting of the engine, and in and through the cylinder on the intake and exhaust
strokes, simply to ensure that those trapping conditions are predicted accurately. Heat transfer
is taking place at every instant of time, to and from all of the gases throughout all of the engine,
and failure to compute it correctly means just that. That is what Chapter 2 is all about.
Secondly, the flow through the valves, ports, and throttles, i.e., all ducting and cylinder
elements, is far from an ideal process. The flow area geometry of poppet valves, as calculated
in this chapter, will be shown to be an optimistic prediction of the actual flow area as the gas
velocity changes or even the direction of that gas flow alters. So, measurements must be made
of the losses accompanying real-world flow regimes, to accompany the gas flow theory of
Chapter 2 to further enhance its simulation accuracy. That is what Chapter 3 is all about.
Thirdly, the gases that flow into an engine are mostly air and those that appear within it
from combustion, and then flow out of it, are defmitely not air. The properties of real gases
differ considerably from air at any temperature, and all gases have properties that change with
temperature. These facts alone deserve detailed study and incorporation within an accurate
simulation of an engine. Combustion itself cannot be treated as a heat transfer process in the
simplistic manner employed in this chapter. Properly carried out, one can predict the local and
mean cylinder temperatures, and the time-varying gas composition, during the burn process so
that one has more accurate information on the composition of the exhaust gas and of the
cylinder state conditions throughout the power cycle. Heat transfer is taking place at every
instant of time, to and from all of the gases present in the cylinder of the engine. Failure to
analyze heat transfer in depth means that we must continue to supply the indices of expansion
and compression to the simulation, rather than vice-versa. All of that theory needs to be incor-
porated into an accurate engine simulation. That is what Chapter 4 is all about.
In Chapter 5, we take stock of what has been learned in Chapters 2-4, and conduct some
modelling of real engines to determine if simulation accuracy has actually improved, as a
result of the ever-increasing sophistication of the thermodynamics and gas dynamics gained in
the interim since Chapter 1.
In Chapter 6, we descend from the lofty heights of the best thermodynamics and gas
dynamics and stoop to empiricism! Here, we search through the simulations to see if there are
simplistic empirical relationships that can be culled, in order to save time in, and give direction
to, the optimization of an engine. The problem with engine simulation is that there are at least
ten times as many data values for the designer to assign as simulation input data as there are
components in the engine. In the search for optimization, it is all too possible to miss the
obvious data "tree" due to the presence of so many other data "trees" within the same engine
129
"wood"! In Chapter 6, we carry out some very useful data tree-thinning. Much as in Sec. 1.8,
we search for empirical simplicity to make more thorough a multi-faceted, complex, optimiza-
tion procedure.
In Chapter 7, we discuss noise. All unsteady gas flow and combustion creates noise. The
perfect silencer, intake or exhaust, produces inflow or outflow characteristics of constant gas
velocity. The quietest engine never fires. The compromise, as always in engineering, must be
sought between the ideal and reality. The unsteady gas flow theory for the ducting is extended
to the propagation of that flow into the atmosphere beyond the ducting terminations so as to
predict the noise it makes at some point in space. This theory is incorporated within the engine
simulation to permit the design of silencers having high gas flow rates with a minimum of
noise creation. The optimum design has a minimum impact on the trapping state conditions at
the beginning ofthe compression stroke, thereby reducing the loss of power by silencing. That
is what Chapter 7 is all about, which brings us back full circle to the start of Chapter 2.
In short, all of the above is what the rest of this book is all about.
References for Chapter 1
1.1 C.L. Cummins, Internal Fire, Society ofAutomotive Engineers, Warrendale, Pa., 1989.
1.2 E.F. Obert, Internal Combustion Engines, 1Oth Printing, International Textbook Com-
pany, Scranton, Pa., 1960.
1.3 C.F. Taylor and E.S. Taylor, The Internal Combustion Engine, International Textbook
Company, Scranton, Pa., 1962.
1.4 C.F. Caunter, Motor Cycles, a Technical History, Science Museum, London, HMSO,
1970.
1.5 H.S. Ricardo, The Pattern ofMy Life, Constable, London, 1968.
1.6 H.S. Ricardo, The High-Speed Internal-Combustion Engine, 4th Edition, Blackie,
London, 1953.
1.7 P.E. Irving, Tuningfor Speed, 3rd Edition, Temple Press Books, London, 1956.
1.8 V. Willoughby, Classic Motorcycles, Hamlyn, London, 1975.
1.9 G.P. Blair, Design and Simulation of Two-Stroke Engines, R- 161, Society ofAutomo-
tive Engineers, Warrendale, Pa., 1996.
1.10 R.C. Cross, "Experiments with Internal Combustion Engines," Chairman's Address,
Proc.I.Mech.E. (Automobile Division), p. 1, 1957-58.
1.11 M.C.I. Hunter, Rotary Valve Engines, John Wiley, New York, 1946.
1.12 L.R.C. Lilly, Diesel Engine Reference Book, Butterworths, London, 1984.
1.13 W.L. Brown, "Methods for Evaluating Requirements and Errors in Cylinder Pressure
Measurement," SAE International Congress, Detroit, SAE Paper No. 670008, Society
of Automotive Engineers, Warrendale, Pa., 1967.
1.14 SAE J1349, Engine Power Test Code, Spark Ignition and Diesel, June 1985.
1.15 Automotive Engineering, Society of Automotive Engineers, Warrendale, Pa., p. 81,
December 1997.
1.16 G.P. Blair, H.B. Lau, A. Cartwright, B.D. Raghunathan, and D.O. Mackey, "Coeffi-
cients of Discharge at the Apertures of Engines," SAE Intemational Off-Highway
Meeting, Milwaukee, Wisc., September 1995, SAE Paper No. 952138, pp. 71-85,
Society of Automotive Engineers, Warrendale, Pa.
130
1.17 G.P. Blair, F.M. Drouin, "The Relationship between Discharge Coefficients and the
Accuracy of Engine Simulation," SAE Motorsports Engineering Conference and
Exposition, Dearbom Mich., December 8-10, 1996, SAE paper No. 962527, Society
of Automotive Engineers, Warrendale, Pa.
1.18 G.P. Blair, D. McBurney, P. McDonald, P. McKernan, and R. Fleck, "Some Funda-
mental Aspects of the Discharge Coefficients of Cylinder Porting and Ducting
Restrictions," Society of Automotive Engineers, Intemational Congress, Detroit, Mich.,
February 1998, SAE Paper No. 980764, Society ofAutomotive Engineers, Warrendale, Pa.
1.19 FY Chen, Mechanics and Design of Cam Mechanisms, Pergamon, Oxford, 1982.
1.20 Cams and Cam Mechanisms (Editor, J.R. Jones), I.Mech.E. Conference, Liverpool
Polytechnic, 1974.
1.21 S. Molian, The Design of Cam Mechanisms and Linkages, Constable, London, 1968.
1.22 G.P. Blair, "Correlation of Measured and Calculated Performance Characteristics of
Motorcycle Engines," Funfe Zweiradtagung, Technische Universitiit, Graz, Austria,
22-23 April 1993, pp. 5-16.
1.23 SAE J604, Engine Terminology and Nomenclature, June, 1995.
1.24 R.S. Benson, N.D. Whitehouse, Internal Combustion Engines, Volumes 1 and 2,
Pergamon, Oxford, 1979.
1.25 J.B. Heywood, Internal Combustion Engines Fundamentals, McGraw-Hill, New York,
1988.
1.26 SAE J1349, Engine Power Test Code, Spark-Ignition and Diesel, June 1995.
1.27 ISO 3046, Reciprocating Intemal Combustion Engines: Performance-Parts 1, 2, and
3, International Standards Organization, 1981.
1.28 SAE J1088, Test Procedure for the Measurement of Exhaust Emissions from Small
Utility Engines, February 1993.
1.29 BS 1042, Fluid Flow in Closed Conduits, British Standards Institution, 1981.
1.30 G.J. Van Wylen and R.E. Sonntag, Fundamentals of Classical Thermodynamics, SI
Version 2e, Wiley, New York, 1976.
1.31 W. Watson, "On the Thermal and Combustion Efficiency of a Four-Cylinder Petrol
Motor," Proc. I. Auto. E., Vol. 2, p. 387, 1908-1909.
1.32 E.W. Huber, "Measuring the Trapping Efficiency of Internal Combustion Engines
Through Continuous Exhaust Gas Analysis," SAE International Congress, Detroit,
Mich., February, 1971, SAE Paper No. 710144, Society of Automotive Engineers,
Warrendale, Pa.
1.33 D. Olsen, P. Puzinauskas, and 0. Dautrebande, "Development and Evaluation of Tracer
Gas Methods for Measuring Trapping Efficiency in 4-Stroke Engines," SAE Fuels
and Lubricants Meeting, Dearborn, Mich., May 4-6, 1998, SAE Paper No. 981382,
Society ofAutomotive Engineers, Warrendale, Pa.
1.34 F. J. Laimbock and R. Kirchberger, "Development of a 150cc, 4-Valve CVT Engine
for Future Emission and Noise Limits," SAE International Off-Highway Meeting,
Milwaukee, Wisc., September 1998, SAE Paper No. 982052, Society of Automotive
Engineers, Warrendale, Pa.
1.35 N. Windrum, The Ulster Grand Prix, Blackstaff Press, Belfast, 1979.
1.36 E. Sher, "The Effect of Atmospheric Conditions on the Performance of an Air-Borne
Two-Stroke Spark-Ignition Engine," SAE Paper No. 844962, 1984.
131
Appendix A1.1 Fundamental Thermodynamic Theory for the Closed Cycle

The thermodynamic statements that are set down here are but a precis of a full presenta-
tion to be found in formal undergraduate texts such as those by Van Wylen [1.30] or Heywood
[1.25]. It is expected that the reader will either be fully aware of this subject matter and can
safely ignore these jottings, or will find them useful as a memory jog from student days past,
or will be unaware of basic thermodynamics in which case the reference text by Van Wylen
[1.30] should be studied first. Nevertheless, as this and the following chapters are read and the
theory studied, there will come moments when these brief statements will elucidate a stubborn
theoretical line in a way that a thousand extra words at that juncture might not.
Al.1.l The Equation of State
Gas properties are specified as pressure, p, and temperature, T, in absolute units of Pa and
K, respectively. Pressure is normally measured by a gauge of some type and the units of such
pressure are referred to as a gauge pressure, pg, above the prevailing atmospheric pressure, Pa.
The absolute pressure, p, is the addition ofthe gauge pressure and the atmospheric pressure:
P = Pg + Pa (Al.l)
The pressure may also be referred to as a dimensionless ratio, as a pressure ratio, P,
defined as:
P= p (Al.2)
Po
where the standard reference atmospheric pressure, po, is defined as 101325 Pa, or 1.01325
bar.
A measured temperature will almost certainly be recorded in units such as Celsius, i.e., as
Tc 'C. The absolute temperature, T, in Kelvin units is found from:
T = Tc + 273 (Al.3)
All gases have a gas constant, R, which is found from the universal gas constant, R, which
has a value of 8314.3 J/kg-molK. For air, which has a molecular weight, M, of 29 kg/kg-mol,
the gas constant, R, is found by:
R =-R 8314.3
M
=
29
=287 J/kgK (A1.4)
132
In any volume of gas, V, which is known to be at a pressure, p, and temperature, T, the

mass ofgas contained therein, m, can be found from the equation ofstate. The units of volume
are in strict SI units, which are m3, and the units of mass are kg. The equation of thermody-
namic state is:
pV = mRT (Al.5)
Hence, from Fig. Al. 1 for the two geometries illustrated:
p1V1 = mRT1 P2V2 = mRT2
The density, p, and the specific volume, v, of a gas are reciprocally related to each other.
The units of density and specific volume are kg/m3 and m3/kg, respectively. They are defined
as, and related to each other, by:
V m 1
v= p -p = (Al.6)
m V v
The equation of state can be modified to incorporate density and specific volume as:
p = pRT pv = RT (A1.7)
Hence, in Fig. Al.1:
Pi = pIRT1 P2v2 = RT2

A1.1.2 The First Law of Thermodynamics for a Closed System
The definition of a closed system is that the mass within the system is constant. The vol-
ume, the pressure, and/or the temperature may change, but the mass does not. As far as an
engine is concerned, a closed system is typified by the gas within the cylinder when all valves
or apertures to the cylinder are closed and the assumption made that the piston rings make a
perfect gas seal with the cylinder walls. A typical situation is sketched in Fig. Al.1, where the
piston in the side-valve engine is seen to move from position 1 to position 2, and the result of
what is visibly a compression process is that the volume decreases from V1 to V2, and the
pressure increases from Pi to P2. Work is expended by the piston to accomplish this compres-
sion process. These data on pressure and volume are graphed in Fig. A1.2. If the piston had
been descending in the sketch, it would have been an expansion process instead.
133
INTERNAL ENERGY, Ui INTERNAL ENERGY, U2

TEMPERATURE, T1 TEMPERATURE, T2
PRESSURE, Pi PRESSURE, P2
VOLUME, V1 VOLUME, V2
roces - - -
- - - so e
Fig. AJ.J A closed sytem piProcess in a four-stroke engine.
J .a - o
i a: -Pi A
V2
a. P2 2
ui
al)
w
- P1
VOLUME, V
Fig. A1.2 The p-V changesfor a closed system process.
134
If the gas within a closed system experiences a work process, SW, it is observed by a
change of volume, dV, at pressure, p, and can be evaluated using the thinking behind
Eq. 1.6.26 as:
8W = pdV (A1.8)
The first law of thermodynamics for a closed system states that any changes of heat trans-
fer, 8Q, and internal energy, dU, are related to any possible work change, 8W, by:
8Q = dU + SW (Al.9)
Actually the above Eq. Al .9 is more completely stated as:
8Q = [dU+dKE +dPE +d?]sysm + 8W
where the term in the square brackets details the possible changes of system energy that could
take place. However, because a closed system containing gas rarely has any significant con-
tent of kinetic energy (KE), or potential energy (PE), or some unknown such as magnetohy-
drodynamic energy, much less changes of them, i.e., dKE, etc., then the norm is that only the
change of internal energy, dU, is ofreal significance.
Eq. Al.9 introduces a sign convention in engineering thermodynamics whereby heat into
a system, and work out ofa system, are defined as being numerically positive. The corollary is
that the reverse direction, i.e., heat out of a system, and work into a system, are defined as
being numerically negative.
The first law of thermodynamics for a closed system also states that cyclic processes can
be evaluated directly, where the definition of a thermodynamic cycle is a series of processes
that culminate in the initial thermodynamic state being restored:
fQ = f8W (Al.lO)
which means that: dU = 0 (Al.1)
The defmition of a thermodynamic cycle should be noted carefully, for in the four-stroke
cycle engine there are two mechanical cycles executed before the actual initial thermody-
namic state is indexed and the totality of that thermodynamic cycle is completed.
The internal energy, dU, can be evaluated for any given process by:
dU = mCvdT (A1. 12)
135
where Cv is the specific heat at constant volume, in units of J/kgK. The implication is that for
any process, such as that sketched in Fig. Al. 1, where the temperature changes from T1 to T2,
one can calculate the change of internal energy involved as follows:
T2 T2
U2 - U1 = f mCvdT = mCv f dT = mCv(T2 - Ti) (Al.13)
TI Ti
Moving the specific heat term to outside the integral sign means that specific heat has
been designated a constant. Although real gases do not have specific heats that are constants
with temperature, the temperature change would have to be many hundreds of degrees Celsius
for it to become a numerical consideration of importance. Thus, for most calculations specific
heat is considered to be a constant. However, when Chapter 4 is approached, and real combus-
tion is considered theoretically, a different perspective will be required.
Further specific, i.e., per unit mass, definitions for heat transfer, work, and internal energy
may be made:
Specific heat transfer, oq Sq =- Q (A1.14)

m
dU
Specific internal energy, du du = (A 1. 15)
Specific work transfer, Sw w =Sm (A1.16)
Thus, Eq. Al.9 can be redefined by dividing across by the mass, m, to give a specific
formulation of the first law of thermodynamics:
Sq = du +Sw (A1. 17)

You will notice that the differential term in front ofheat and work is a 8 and not a d, as in
front of intemal energy. This is because heat and work are path functions, i.e., the outcome of
any process depends on the path, and intemal energy is a point function, i.e., the outcome is
dependent only on the initial and final states. Mathematically, this is of real significance:
2 2 2 2
jQ=QQ2 JSW=W2JpdV
1
pdU=mC
1 1 1
136
In short, the internal energy can be found by knowing only the initial and final tempera-
tures but, taking work as the example, unless we know the path function connecting p and V
we have no possibility of evaluating the work integral. You can see that in Eq. A1.25, where
the path is defined by the index n. If we do not know that number, we cannot evaluate the work
content.
There are two specific heats to be defined for a gas: the specific heat at constant volume,
Cv, and the specific heat at constant pressure, Cp. The formal defmitions of each are:
du
d(u +pv)dh
dT dT dt (A1.18)
where the variable, h, is specific enthalpy, the definition ofwhich is encapsulated in the above
equation. It transpires that the ratio of these specific heats is a "constant" and is also a property
of the gas. The fact is that specific heats for real gases do vary somewhat, but not hugely, with
temperature. This will be discussed later in the text. The ratio of specific heats, y, is defined as:
Cp (Al.19)
It also transpires that these specific heats are related to the gas constant, R, from the formal
definitions of enthalpy and internal energy, and the state equation as Eq. Al. 7:
dh = CpdT = d(u + pv) = CVdT + d(RT) - CVdT + RdT
hence, Cp = Cv + R (Al.20)
Consequently, combining Eqs. Al.19 and 1.20:
Cv R C =yR (Al.21)
The value of specific heat at constant volume, CV, for air is measured at 718 J/kgK (via
Eq. A1.33; see below). From Eq. A1.20 the specific heat at constant pressure for air, Cp, is
calculated to be 1005 J/kgK, as the gas constant for air, R, is a known quantity at 287 J/kgK.
The ratio of specific heats, y, for air is seen from Eq. Al.19 to be 1.4.
137
A1.1.3 An Adiabatic Work Process in a Closed System

By definition, an adiabatic process has zero heat transfer to, or from, the system across its
boundary:
8Q = 0 (Al.22)
If an adiabatic process occurs and it is as sketched in Fig. Al. 1, where the pressure, tem-
perature, and volume of the closed system change from pi to P2, T1 to T2, and V1 to V2,
respectively, the first law of thermodynamics will now state, from the integration of Eq. Al .9:
2 2
0 = JdU + SW
1 1
With Eq. A1.21, this yields the work transfer, W2:

2
2 2P22-IV
WI= W= -JdU = -mCv(T2 - Ti) P2V2 ( pL 23
Suppose this same adiabatic work process shown could be represented, as in Fig. A1.2, by
a polynomial function connecting pressure and volume with an exponent n, and where the
symbol k is a constant:
pVn =k (Al.24)
This work process could be evaluated by:
2 2 1-n v - p2V2 - pV
J PdV = kV'-ndV ==1~-- k 1-n 1-n2PAl25
1 1
which is found by remembering, from Eq. A1.24, that:
pIVIn = P2V2n = k
138
Chapter I - Introducdon to the Four-Stroke Engine
Comparison of Eqs. A1.25 and Al .23 shows that the polynomial indices, n, and y, must be
identical. In short, in an adiabatic process in a closed system the polynomial index n is the ratio
of specific heats, y. Hence, the work path fimction becomes:
pVy = k (A1.26)
A1.1.4 A Work Process in a Closed System with Heat Transfer
The sketch in Fig. Al.1 shows a work process that could well be taking place with heat
transfer to, or from, the cylinder. Indeed, a simple combustion process could be simulated by
assigning the buming of fuel as a heating process that occurs internally without any mass
transfer, where the cylinder gas may be losing heat to the walls at the same time.
Heat Transfer To, or From, the System
First, let the simpler case of heat transfer to, or from, the cylinder walls be considered.
Eq. Al.9 is integrated again, but this time the polynomial index is n, and n cannot be equal to
y, for the process is no longer adiabatic. Where the index, n, is not equal to y, that index is
known as a polytropic index to distinguish it from y, which is labeled as the isentropic index.
2 2 2
J Q = JdU + JSW (A1.27)
1 1 1
Q2 = mCv(T2 -T) + P2V2 P (A1.28)

1-n(Al28
From the state equation:
p1V1 = mRT1 P2V2 = mRT2 (A1.29)

From the polytropic function:
PVn
pV1 =
Vn
P2V2 =k (A1.30)
In most numerical problems, enough data are known at the commencement of any step
position, 1, to solve the three Eqs. Al.28-Al.30 to evaluate all unknown property data at step
position 2.
For example, if the initial pressure, temperature, and volume, Pl, T1, and V1, the final
volume, V2, and te amount of heat transfer, Q2, are all known, then the final pressure and
temperature, P2 and T2, can be evaluated. In reverse, by measuring cylinder pressures and
139
volumes, it becomes possible to evaluate the amount of heat transfer, Qj2. As shown later in
Chapter 4, this is actually the theoretical basis for the evaluation ofheat release from combus-
tion.
An Adiabatic Process with Internal Heat Transfer
Second, let us use this solution to deal simply with a combustion process that is assumed
to act as an internal heat transfer to an adiabatic system. The adiabatic definition refers to zero
external heat transfer to, or from, the system. If the system is defined as adiabatic, then the
above theory can be used, but the polytropic index, written above as n, reverts to the ideal, i.e.,
y. The amount of internal heat transfer due to combustion becomes the numeric input value for
2
Qi
A1.1.5 Heat Transfer Processes at Constant Volume in the Closed Cycle
Consider the situation in Fig. Al.1, but where the piston does not move. The volume will
stay constant, i.e., V2 equals V1. In such a process all work transfer will be zero:
2
_..~n..
W1,2 -jpdV=0 (Al.31)
However, heat transfer could take place, during which the temperature would rise or fall,
depending on the direction of that heat transfer process. It will be evaluated, as usual, by the
first law of thermodynamics for the closed system, as in the mass specific version of
Eq. Al.17.
6q = du +6w = du + 0 = CvdT (A1.32)

Indeed, this is how the specific heat at constant volume can be measured for any gas by
carrying out heat transfer experiments in a constant-volume, closed system. Integrating Eq.
A1.32 over the process from state conditions 1 to 2 and including the system mass:
2 2
Q2 = J6Q = mCvJ dT = mCv(T2 - T,) (A1.33)
1 1
The system mass is constant and can be found from the state equation at the initial or final
conditions at constant volume, where V1 equals V2:
m= piv1 = P2V2 (A1.34)
140
This process is precisely that specified in the discussion of the ideal Otto cycle in Sec.
1.6.8 as item (a) for the combustion process at tdc and as item (d) for the exhaust heat rejection
process at bdc. These two processes are evaluated by the use of Eqs. A1.33 and A 1.34. For
example, the heat addition process at tdc is found by equating the heat transfer, Q2, to the heat
released by the fuel during the ideal constant volume process and solving for the unknown
values of pressure and temperature at the end of the process, P2 and T2. More fundamental
theory, which uses these thoughts regarding the ideal Otto cycle, is to be found in this appen-
dix in Sec. Al. 1.8.
A1.1.6 Direction of Heat Transfer in a Polytropic Process in a Closed Cycle
Consider the use of the first law for the analysis of a closed system where both polytropic
work transfer and heat transfer are taking place. Using a modified version of the solution
achieved in Eq. A1.28 for this type of process:
First Law in the specific format:
Sq = du + Sw
From Eq. A1.28:
&1 vdT+ RT(Al.35)

8q=C - n(A.5
Rearranging: Sq = dT(Cv + n (Al.36)
Inserting Cv from Eq. A1.21:
q =dT(y- +j- J (Al.37)
Rearranging: Sq n1RdT(y
( - (Al.38)
A Compression Process
(i) In a compression process, the temperature, dT, will rise, i.e., dT is positive. Note that
the values of the ratios of n and y are always greater than unity.
141
If heat is added to the system, i.e., if Sq is positive, then from Eq. A1.38:
n>y (A 1.39)
If heat is lost from the system, i.e., if Sq is negative, then from Eq. A1.38:
(Al.40)
An Expansion Process
(ii) In an expansion process, the temperature, dT, will fall, i.e., dT is negative.
If heat is added to the system, i.e., if Sq is positive, then from Eq. A1.38:
n< y (Al.41)
If heat is lost from the system, i.e., if Sq is negative, then from Eq. A1.38:
n > y(Al.42)
It will be observed, supporting the contentions of Eqs. Al.23-Al.25, that if n equals y,
then the heat transfer is always zero and the process must be adiabatic as a consequence.
A1.1.7 Gas Property Relationships in Adiabatic and Polytropic Processes
Consider a polytropic process proceeding from thermodynamic state 1 to state 2, as sketched
in Fig. Al.1.
From the state equation:
p1V1 = mRT1 P2V2 = mRT2 (Al.43)
Dividing these two equations:
p1v1 = mRT1 T_
=
(Al.44)
P2V2 mRT2 T2
With specific volume instead:
p1v1 - mRT1 T,
=
(Al.45)
P2v2 mRT2 T2
142
With density instead:
Pi - pIRT1 - pIT,
(A 1.46)
P2 P2RT2 P2T2
From the polytropic equations:
pVn = k (Al.47)
which also can be expressed as: pvn = k (Al.48)
or as: p= kpn (Al.49)
Integrating between any two thermodynamic states 1 and 2, Eqs. A1.46-A1.48 become:
Pi =
()-n
(Al.50)
P2 tV2)
-n
Pi =
(Al.51)
P2 Jv2)
n
£L =[A (Al.52)
P2 %P2)
Incorporating the temperature relationships in Eqs. Al .44-A1.46 will give modified ver-
sions of Eqs. Al.50-Al.52 as:
1 n-I
Plvi = Pi (Pi n-
= p1 n = I1 (A1.53)
P2V2 P2 P2 tP2) T2
n
or: P2 = (T2 n-I (Al.54)
P2 iT2 )
143
VV-n VI VI l-n T
(Al.55)
or:
P2V2 V2 ) V2 V2) T2
or: VI vl jTiJl-n (Al.56)

V2 V2 tT2)
or: Pi = (TI n-I (Al .57)

P2 T2)
It should be made clear, if the process is isentropic and adiabatic where the polynomial
index is y, rather than n as in a polytropic case, then the index y will replace the index n in Eqs.
Al .43-A1.57.
A1.1.8 The Thermal Efficiency of the Ideal Otto Cycle
The temperature-volume characteristics of this cycle are shown in Fig. A1.3. Sec. 1.6.9
describes how there are only two heat transfer processes in this cycle. The first is an intemal
heat addition process, i.e., to simulate combustion at tdc, taking place at constant volume as an
adiabatic process from states labeled as 1 to 2 in Fig. A1.3.
4500 -
4000 - HEAT ADDITION

3500 - AT CONSTANT VOLUME
L
3000
I-
2500 -
wa.2000 -
4
w 1500
H-
1000 -
500 -
1
0* i
I *I U * *
0 100 200 300 400 500 600

Fig. AJ.3 The temperature-volume characteristics of the ideal Otto cycle.
144
The second is an internal heat rejection process, i.e., exhaust but without any mass flow at
bdc, taking place at constant volume as an adiabatic process from states labeled as 4 to 1 in
Fig. A1.3. The initial state 1 is so indexed and the thermodynamic cycle is completed. The
other two processes labeled as 2-3 and 3-4 are adiabatic and isentropic, and are compression
and expansion processes, respectively. By definition from Sec. Al .1.5, these processes have
zero heat transfer characteristics. The thermal efficiency of this cycle is defined in Eq. 1.6.24
as
net work output

heat input (Al .58)
There being only one heat input, and two work processes, reduces to:
W2 +W34
7t = .
- 3-
Q2
(Al.59)
From Eqs. 1.23 and 1.33 for such work and heat processes, Eq. A1.59 becomes:
-mCv T) -mCv(T4 - T3)

-
tIt = mvT 2 (Al .60)
mcv -T2)
T4-T1 ILL!T L I
Rearranging: (Al.61)
T3 - T2 T2 I.
T3 lIJ
_
However, from Eq. A1.55:
(VL ) T, V43) T4
= =
(Al.62)
YV2) T2 T3
Compression ratio, CR, is defined as:
CR = VI = V4 (Al.63)
V2 V3
145
Hence, Eq. A1.62 shows:
T2 = T4
(Al.64)
Rearranging: T2 T1 (Al.65)
Using Eqs. A1.62, A1.63, andA1.65, Eq. A1.61 becomes:
nt = -l
1 = I - CR1-' = 1- 1 (Al.66)
T2 CR'Y-1
A1.1.9 A Constant Pressure Process with Work and Heat Transfer in a Closed System
In the next section, a discussion will be presented on a thermodynamic process in a closed
system, taking place at constant pressure, where there is both work and heat transfer. Let us
examine the basic thermodynamics of it here. The first law of thermodynamics for a closed
system can be applied to this case. From Eq. Al.17:
Sq = du + Sw
From Sec. Al.1.2:
Sq = CVdT + pdv (Al.67)

Because pressure is a constant, differentiating Eq. Al.7 yields:
d(pv) = d(RT)
The left side of the equation becomes:
d(pv) = pdv + vdp = pdv

The right side of the equation becomes:
d(RT) = RdT
Hence, the work term is:
Sw = pdv = RdT (A1.68)
146
and heat transfer is:
Sq = CvdT + RdT = (Cv + R)dT = CpdT (A1.69)

In short, the heat transfer in a constant pressure process in a closed system can be logically
evaluated using the specific heat at constant pressure and the initial and final state conditions
of temperature.
A1.1.10 The Thermal Efficiency of the Ideal Diesel Cycle
The pressure-volume and temperature-volume characteristics of the ideal version of this
cycle are shown in Figs. A1.4 and Al.5.
It is seen that, compared to the Otto cycle in Fig. A1.3, there is an extra heat transfer
process in this cycle. The total amount ofheat to be added at the tdc location is carried out in
two internal heat transfer processes: a constant-volume process followed by a constant-
pressure process. In many texts [1.2, 1.3], the ideal Diesel cycle is defined as having all ofthe
heat addition take place at constant pressure only, i.e., the constant-volume segment does not
exist. The version of the ideal Diesel cycle that has both constant-volume and constant-
pressure heat addition segments is often referred to as the Dual cycle [1.2]. This is semantics,
as both are ideal versions of a cycle attempting to simulate the thermodynamics in a real diesel
engine.
Apart from the mechanism of internal heat addition, the cycle is identical to the Otto
cycle. The initial state 1 is indexed and the thermodynamic cycle commenced. Adiabatic isen-
tropic compression takes place from state 1 to state 2. Constant volume heat addition takes
place internally and adiabatically from state 2 to state 3. Constant pressure heat addition takes
place internally and adiabatically from state 3 to state 4. Constant volume heat rejection in this
closed system, to simulate the real exhaust process in an open system, takes place adiabati-
cally as an internal heat transfer from state 4 to state 5.
The thermal efficiency of this, or any, thermodynamic cycle is defined in Eq. 1.6.24 as
net work output (A1.70)

= heat input
There being two heat input, and three work processes, this reduces to:
wI + W3 + WI
t 3
Q2+Q3 =Q4 (A1.71)
147
140 HEAT ADDITION AT:

L- CONSTANT VOLUME FROM 2-3
120 CONSTANT PRESSURE FROM 3-4
co
LIJ
100
CD
Ca
ci)
CO
w 80
a:
0L
a: 60 \ IDEAL DIESEL CYCLE
w
ci
z 40
0i
20 5
1
0
0 200 400 600 800 1000
Fig. Al.4 The pressure-volume characteristics of the ideal Diesel cycle.
2000 4
HEAT INPUT AT CONSTANT PRESSURE
cc
lll 1800
a:I- 1600 HEAT INPUT AT CONSTANT VOLUME
cc 1400
w 1200
a.
w 1000
a: 800 5
w
600
z
400
0i 1
200
0- 6 I v s a I v n 9 5 "'"
0 200 400 600 800 1000
Fig. Al.5 The temperature-volume characteristics of the ideal Diesel cycle.
148
Chapter I - Introducdion to the Four-Stroke Engine
From Eqs. 1.23, 1.33, and Al.69 for such work and heat processes, Eq. A1.71 becomes:
-mCv(T2 - Ti) + mR(T4 - T3) - mCv(T5 - T4)

mCv(T3 - T2) + mCp(T4 T3)
Rearranging: -(T= - Ti ) XT4 3) -(T5 T4) (Al.73)
During any given analysis, the total heat, QR, to be added by heat transfer to simulate
combustion will be a known numeric quantity. No numeric solution for thermal efficiency, or
indeed any other cyclic parameter, is possible unless a decision is made regarding the bum
proportion, kcv, of the total heat transfer to be added during the constant volume segment at
tdc. This burn proportion, kcv is defined as follows:
3
- heat added at constant volume Q2
total heat added QR Q3 + Q(A1.74)
If the value of the burn proportion, kcv, is unity, then the solution reverts to that for the
ideal Otto cycle in Sec. A1. 1.8. If the value of kcv is zero, then the result is an ideal Diesel
cycle with heat added only at constant pressure, and falls under the classic definition of the
ideal Diesel cycle. Any value of kcv between 0 and 1 defines it, as Obert [1.2] would have
done, as the Dual cycle. Because we are working with a Diesel cycle, it is assumed here that
the value of kcv must be other than unity.
Irrespective of the value of bum proportion used in any given circumstance, the determi-
nation of cyclic performance parameters requires, as usual, determination of the state condi-
tions of pressure, temperature, and volume throughout the cycle. The initial state conditions
must be defined, as must the total amount ofheat to be internally added as simulated combus-
tion. The swept volume and compression ratio must also be defined, which then provides the
volumes at bdc and tdc, i.e., V1 and V2, and indeed V3 as well by inference.
Because all processes in the engine are adiabatic, the first compression process from state
1 to 2, and the heat addition at constant volume can be analyzed as for the Otto cycle. Thus, to
obtain T3 the following equations are solved:
By definition:
Q2 = kcvQR = (1-kkv)QR (A1.76)
149
From Eq. A1.33
3= mICV(T3 - T2) (A1.75)

'From Eq; A1.34:
P3 = P2 T (A1.76)
T2
From Eq. A-1.69
4= mICp(T4 T3) (A1.77)

By definition:
P4 = P3 (A1.78)
Hence, from Eq. A1.5
V T3 (A1.79)
From Eq. A1.55:
T5 =
T4(-V)J (A1.80)
From Eq. A1.26:
,V48
= P4 (A1.81)
From Eq. A1.23:
W -mICv(T2 - Ti) P2V2 P1V- = (A1.82)
150
From Eq. A1.68:
3 = mlR(T4 - T3) = p3(V4 - V3) (A1.83)
From Eq. A1.23:
5= -m CV(T5 T ) - P5 - P4V4 (A1.84)
From Eq. A1.33:
= m1Cv(T1 - Ts) (A1.85)

From Eq. 1.6.31:
2 4
imep = Wi
WI + 3+W4 (A1.86)
VV
The thermal efficiency is found by using whichever of Eqs. A1.71-Al.73 is the most con-
venient, having evaluated all of the state conditions for points 1-5 and the work, heat transfer,
and energy processes which connecting them.
The above analysis includes both constant-volume and constant-pressure segments for the
heat transfer process which simulate combustion. In the event that a particular simulation does
not call for any constant-volume combustion, then the cycle analysis is somewhat simplified.
Ideal Diesel Cycle with Constant-Pressure Heat Transfer Only
The following simplifications apply for insertion into Eqs. 1.70-Al.86
kC =°
cv Q32 =0
Q2Q3=QR4Q = QR(A1 .87)
P3 = P2 T3 = T2 U3 = U2 (A1.88)
In this case, it is very simple to show that the equation for thermal efficiency, Eq. A1.73,
can be simplified as:
n 1 (1(T5-TI
- HT5-I(Al.89)
151
Today, because numerical analysis by computer cares little for the algebraic simplifica-
tions that graced yesterday's textbooks [1.2], it hardly seems worth the effort to set it down
here. Back when in my student days, and armed only with a slide rule to get it reduced to
numbers, numerical analysis was a major issue. Trying to remember it for the (one and only)
annual examination on the subject was even worse!
It is also possible to show [1.2] that this version of the ideal Diesel cycle has a thennal
efficiency that is less than the ideal Otto cycle at equal compression ratios. Because any diesel
engine operates at a compression ratio level that is about twice that of its Otto engine equiva-
lent, this gratuitous information is of no real significance.
152
Chapter 2
Gas Flow through Four-Stroke Engines

2.0 Introduction
The gas flow processes into, through, and out of an engine are all unsteady. Unsteady gas
flow is defined as that in which the pressure, temperature, and gas particle velocity in a duct
are variable with time. In the case ofexhaust flow, the unsteady gas flow behavior is produced
because the cylinder pressure falls with the rapid opening of the exhaust valve or valves. This
gives an exhaust pipe pressure that changes with time. In the case of induction flow into the
cylinder through an intake valve whose area changes with time, the intake pipe pressure alters
because the cylinder pressure is affected by the piston motion causing volumetric change
within that space.
To illustrate the dramatic variations of pressure wave and particle motion caused by
unsteady flow compared to that caused by steady flow, a series of photographs obtained by
Sam Coates and me during research at QUB [7.2] is shown in Plates 2.1 to 2.4. These photo-
graphs were obtained using the Schlieren method [7.17], an optical means of observing the
variation of the refractive index of a gas with its density. Each photograph was taken with an
electronic flash duration of 1.5 ts and the view observed is around the termination of a 28-mm
diameter exhaust pipe to the atmosphere. The exhaust pulsations occurred at a frequency of
1000 per minute. The first photograph, Plate 2. 1, shows the front of an exhaust pulse about to
enter the atmosphere. Note that this is a plane front, and its propagation within the pipe up to
the pipe termination is clearly one-dimensional. The next photograph, Plate 2.2, shows the
propagation of the pressure wave into the atmosphere in a three-dimensional fashion with a
spherical front being formed. The beginn ng of rotational movement of the gas particles at the
pipe edges is now evident. The third photograph, Plate 2.3, shows the spherical wave front
fully formed and the particles being impelled into the atmosphere in the form of a toroidal
vortex, or a spinning donut of gas particles or "smoke ring." This propagating pressure wave
front arrives at the human eardrum, which deflects it, with the nervous system reporting it as
"noise"V to the brain. The final photograph of the series, Plate 2.4, shows that the propagation
ofthe pressure wave front has now passed beyond the frame of the photograph, but the toroi-
dal vortex of gas particles is proceeding downstream with considerable turbulence. Indeed, the
flow through the eye of the vortex is so violent that a new acoustic pressure wave front is
forming in front of that vortex. The noise that emanates from these pressure pulsations is
composed of the basic pressure front propagation and also of the turbulence of the fluid
153
Plate 2.1 Schlieren picture of an exhaust pulse at the termination ofa pipe.
motion in the vortex. Further discussion on the noise aspects of this flow is given in
Chapter 7. I have always found this series of photographs to be particularly illuminating.
When I was a schoolboy on a farm in Co. Antim too many years ago, the milking machines
were driven by a single-cylinder four-stroke Lister diesel engine with a long, straight exhaust
pipe and, on a frosty winter's morning, it would blow a "smoke ring" from the exhaust on
start-up. That schoolboy used to wonder how it was possible; I now know.
Because the performance characteristics of an engine are significantly controlled by this
unsteady gas motion, it behooves the designer of engines to understand this flow mechanism
thoroughly. This is true for all engines, whether they are destined to be a 2 hp lawnmower
engine or a 1000 hp offshore boat racing engine. A simple example will suffice to illustrate the
point. If one were to remove the tuned exhaust pipe from a single-cylinder racing engine while
it was running at peak power output, the pipe being an "empty" piece of fabricated sheet
metal, the engine power output would fall by at some 20% at that engine speed. The tuned
exhaust pipe harnesses the pressure wave motion of the exhaust process to extract a greater
mass ofthe exhaust gas from the cylinder during the exhaust stroke and initiate the induction
process during the valve overlap period. Without it, the engine would only be able to inhale
about 80% as much fresh air and fuel into the cylinder. To design such exhaust systems, and
the engines that will take advantage of them, it is necessary to have a good understanding of
the mechanism of unsteady gas flow. For the more serious student interested in a more
154
Chapter 2 - Gas Flow through Four-Stroke Engines
Plate 2.2 The exhaust pulse front propagates into the atmosphere.
in-depth treatment of the subject of unsteady gas dynamics, the series of lectures given by the
late Prof. F. K. Bannister of the University of Birmingham [2.2] is an excellent introduction to
the topic; so too is the book by Annand and Roe [3.12] and the books by Rudinger [2.3] and
Benson [2.4]. The references cited in this chapter will give even greater depth to that study.
This chapter explains the fudamental characteristics of unsteady gas flow in the intake
and exhaust ducts of reciprocating engines. Such fundamental theory is just as applicable to
two-stroke engines as it is to four-stroke engines, although the bias of the discussion will
naturally be toward the four-stroke engine. Throughout the chapter, the relevance of each
unsteady gas flow topic is discussed in the context of the design of tuned exhaust and intake
systems for four-stroke engines.
2.1 Motion of Pressure Waves in a Pipe
2.1.1 Nomenclaturefor Pressure Waves
The motion of pressure waves of small amplitude is already familiar to us through our
experience with acoustic waves, or sound. Some ofour experience with sound waves is help-
fil in understanding the fundamental nature ofthe flow ofthe much larger-amplitude waves to
be found in engine ducts. As illustrated in Fig. 2.1, pressure waves, and sound waves, are of
two types. They are either compression waves or expansion waves. In both Fig. 2. 1(a) and (b),
the undisturbed pressure and temperature in the pipe ahead of the pressure wave are po and To,
respectively.
155
Design and Simulation ofFourStroke Engines
Plate 2.3 Further pulse propagation followed by the toroidal vortex ofgas particles.
The compression wave in the pipe is shown in Fig. 2.1(a) and the expansion wave in Fig.
2. 1(b). Both waves are propagating toward the right in the diagram. At a point on the compres-
sion wave, the pressure is Pe, where Pe is greater than po, and the wave is being propagated at
a velocity a,. It is also moving gas particles at a gas particle velocity ce, in the same direction
of propagation as the wave. At a point on the expansion wave, the pressure is pi, where pi is
less than po, and the wave is being propagated at a velocity ai. It is also moving gas particles
at a gas particle velocity ci, but in a direction opposite to the direction of propagation of the
wave.
At this point, our experience with sound waves helps us to understand the physical nature
of the statements made in the preceding paragraph. Imagine standing several meters away
from another person, Fred. Fred produces a sharp exhalation of breath, for example, he says
"boo" somewhat loudly. He does this by raising his lung pressure above the atmospheric pres-
sure due to a muscular reduction of his lung volume. The compression pressure wave pro-
duced, albeit of small amplitude, leaves his mouth and is propagated at the local acoustic
velocity, or speed of sound, to your ear. The speed of sound involved is on the order of
350 m/s. The gas particles comprising the "boo" leaving Fred's mouth have a much lower
velocity, probably on the order of 1 m/s. However, the gas particle velocity is in the same
direction as the propagation of the compression pressure wave, i.e., toward your ear. Contrast
this simple experiment with a second test. Imagine that Fred now produces a sharp inhalation
of breath. This he accomplishes by expanding his lung volume so that his lung pressure falls
156
Plate 2.4 The toroidal vortex ofgas particles proceeds into the atmosphere.
IZ4
ae
A d AT
a
I11
1tT
a
d
(a) compresion prsure wave
(b) expansion e wave
Fig. 2.1 Pressure wave nomenclature.
157
sharply below the atmospheric pressure. The resulting "u... uh" you hear is caused by the
expansion pressure wave leaving Fred's mouth and propagating toward your ear at the local
acoustic velocity. In short, the direction of propagation is the same as before with the compres-
sion wave "'boo," and the propagation velocity is, for all intents and purposes, identical. How-
ever, because the gas particles manifestly entered Fred's mouth with the creation of this
expansion wave, the gas particle velocity is clearly in the direction opposite to the expansion
wave propagation.
It is obvious that exhaust pulses resulting from cylinder blowdown, when the exhaust
valve opens, fall under the category of compression waves, whereas the waves generated by
the rapidly falling cylinder pressure during induction with the intake valve open are expansion
waves. However, as will be seen from the following sections, both expansion and compression
waves appear in the inlet and the exhaust system.
NOTE: As in most technologies, other jargon is used in the literature to describe compres-
sion and expansion waves. Compression waves are variously called "exhaust pulses," "com-
pression pulses," or "ramming waves." Expansion waves are often described as "suction pulses,"
"sub-atmospheric pulses," "rarefaction waves," or "intake pulses."
2.1.2 Propagation Velocities ofAcoustic Pressure Waves
As already pointed out, acoustic pressure waves are pressure waves with small pressure
amplitudes. Let dp be the pressure difference from atmospheric pressure, i.e., (Pe-PO) or
(Po-Pi), for the compression or expansion wave, respectively. The value of dp for Fred's "boo"
would be on the order of 0.2 Pa. The pressure ratio, P, for any pressure wave is defined as the
pressure, p. at any point on the wave under consideration divided by the undisturbed pressure,
po, more commonly called the reference pressure. This is normally the standard reference
pressure, 101,325 Pa or 1.01325 bar. Here, the pressure ratio for Fred's "boo" would be:
p p 101325.2 1.000002
Po 101325
For the loudest of acoustic sounds, say, a rifle shot at about 200 mm from the human ear,
dp could be 2000 Pa and the pressure ratio would be 1.02. That such very loud sounds are still
small in terms of pressure wave can be gauged from the fact that a typical exhaust pulse in an
engine exhaust pipe has a pressure ratio of about 1.5.
According to Eamshaw [2.1], the velocity of a sound wave in air is given by a0, where:
ao= iyRT (2.1.1)
or ao_ (2.1.2)
P0
158
The value denoted by y is the ratio of specific heats for air. To is the reference temperature
and ro is the reference density, which are related to the reference pressure, po, by the state
equation:
Po = poRTo (2.1.3)
For sound waves in air, po, To, and ro are the values of the atmospheric pressure, tempera-
ture, and density, respectively, and R is the gas constant for the particular gas involved. Con-
sult Appendix A1 .1 for more information on these symbols and their definitions.
2.1.3 Propagation and Particle Velocities ofFiniteAAmplitude Waves
Particle Velocity
Any pressure wave with a pressure ratio greater than that of an acoustic wave is called a
wave of fnite amplitude. Earnshaw [2. 1] showed that the gas particle velocity associated with
a wave of fmnite amplitude is given by c, where:
F 1
2 Vp2
c -ao1 [PPo) 1 (2.1.4)
y-l
Bannister's [2.2] derivation of this equation is explained with great clarity and is pre-
sented in Appendix A2. 1. Within the equation, shorthand parameters can be employed which
simplify the understanding of much ofthe further analysis. The symbol P is referred to as the
pressure ratio ofa point on a wave ofabsolute pressure, p. The notation of X is known as the
pressure amplitude ratio, and G represents various functions ofy, which is the ratio ofspecific
heats for the particular gas involved. These are set down as:
Pressure ratio P -
Po
Pressure amplitude ratio X= (2.1.5)
Incorporation of the above shorthand notation into Eq. 2.1.4 gives:

2
c=
y -l
lao(X-1) (2.1.6)
159
If the gas in which this pressure wave is propagating has the properties of air, then these
properties are:
Gas constant R=287 J/kgK
Ratio of specific heats y =1.4
Specific heat at constant pressure CPPy-1 = 1005 J/kJK

y-
Specific heat at constant volume R

CV= --= 718 J/kgK
Various fimctions ofthe ratio of specific heats, G5, G7, etc., which are useful as shorthand
notation in many gas dynamic equations in this theoretical area, are given below. The logic of
the notation for G is that the value of G5 for air is 5, G7 for air is 7, etc.
G3 = 4 -2 =3forair, wherey = 1.4

y -1
G4 = 3-y - 4forair, where y = 1.4

y-1I
G5 = .l = 5 forair, wherey = 1.4
G6= + I= 6forair, wherey = 1.4
G7 = 2y - 7 forair, wherey = 1.4

y-1I
I
G17 = y I =
-
for air, wherey = 1.4
(335 = - = 3.5 for air, where y = 1.4

y-1
67 2y 6 for air, where y = 1.4
160
This useful notation simplifies analysis in gas dynamics, particularly because the equa-
tions are additive or subtractive with numbers, thus:
For example
=G5
G4=G5 -1= 2 -1= 2-y+1 1-y
l-l-y-l
Y-1 Y-1 Y-1
or G7=G5 +2 or G3=Gs -2 or G6=G3+3

However, it should be noted in applications of such functions that they are generally nei-
ther additive nor operable, thus:
G7*G4+G3 and G 26
Gas mixtures are commonplace within engines. Air itself is a mixture-basically of oxy-
gen and nitrogen. Exhaust gas is principally composed of carbon monoxide, carbon dioxide,
steam, and nitrogen. Furthermore, the properties of gases are complex functions of tempera-
ture. A more detailed discussion ofthis topic is therefore necessary, and is given in Sec. 2.1.6.
Ifthe gas properties are assumed to be as for air with the properties above, then Eq. 2.1.4
for the gas particle velocity reduces to the following:
2
c = - ao(X -1) = G5ao(X -1) = 5aO(X - l) (2.1.7)
y-l
where x=(J (JG) {p-} (2.1.8)
Propagation Velocity
The propagation velocity at any point on a wave, where the pressure is p and the tempera-
ture is T, is like that of a small acoustic wave moving at the local acoustic velocity under those
conditions, but on top of gas particles which are already moving. Therefore, the absolute propa-
gation velocity of any point on a wave is the sum of the local acoustic velocity and the local
gas particle velocity. The propagation velocity ofany point on a finite amplitude wave is given
by a, as:
a=a+c (2.1.9)
161
where a is the local acoustic velocity at the elevated pressure and temperature of the wave
point, p and T.
Acoustic velocity, a, is given by Eamshaw [2.1] from Eq. 2.1.1 as:
a=yT (2.1.10)
Assuming the change in state conditions from po and To to p and T to be isentropic, and
using the theory as given in AppendixAl.1, Sec. A1.1.7 we have:
a-1
To Po
y-1
a- -- P017=X (2.1.12)
ao To tPo°
Hence, the absolute propagation velocity, a, defined by Eq. 2.1.9, is given by the expres-
sions for a and c given by Eqs. 2.1.6 and 2.1.12:
a = aoX+ -
y-1ao(X-
1) ao
=
(2.1.13)
LY1P -
In terms of the G functions already defined,
a=a0[G6X-G5] (2.1.14)
If the properties of air are assumed for the gas, then this reduces to:
a=ao[6X-5] (2.1.15)
162
The density, p, at any point on a wave of pressure p is found from an extension of the
isentropic relationships in Eqs. 2.1.11 and 2.1.14 as follows, using the isentropic theory of
Appendix A1.1, Sec. Al1.1.7:
1 2
p p Y
-
= XY-1 = XG5 (2.1.16)
Po . PO)
For air, where y is 1.4, the density, p, at a pressure p, on the wave translates to:
p = poX5 (2.1.17)
2.1.4 Propagation and Particle Velocities ofFinite Amplitude Waves in Air

From Eqs. 2.1.4 and 2.1.15, the propagation velocities of finite-amplitude waves in air in
a pipe are calculated by the following equations:
Propagation velocity a= ao6X-51 (2.1.18)
Particle velocity c=5ao(X-1) (2.1.19)
Pressure amplitude ratio X= j =P (2.1.20)
The reference conditions of acoustic velocity and density are found as follows:
Reference acoustic velocity ao= 1.4x287xT0oM/s (2.1.21)
Reference density P =
287xT kg/m3 (2.1.22)
It is interesting that these equations corroborate the experment that we conducted with
our imaginations regarding Fred's lung-generated compression and expansion waves.
163
Fig. 2.1 shows compression and expansion waves. Let us assume that the undisturbed
pressure and temperature in both cases are at standard atmospheric conditions. In other words,
po and To are 101,325 Pa and 20°C, or 293 K, respectively. The reference acoustic velocity, ao,
and reference density, po, are, from Eqs. 2.1.1 and 2.1.3 or Eqs. 2.1.21 and 2.1.22:
ao = 1.4x287x293 = 343.1 m/s
Po -1135=1. 2049 kg/in3

287 x 293
Let us assume that the pressure ratio, Pei of a point on the compression wave is 1.2 and that
of a point on the expansion wave, Pi, is 0.8. In other words, the compression wave has a
pressure differential as much above the reference pressure as the expansion wave is below it.
Let us also assume that the pipe has a diameter, d, of 25 mm.
The Compression Wave
First, consider the compression wave of pressure Pe:
Pc = Pe x Po = 1.2 x 101325= 121590 Pa

The pressure amplitude ratio, Xe, is calculated as:
x = 1.27 = 1.02639
Therefore, the propagation and particle velocities, ae and ce, are found from:
a, =343.llx(6x1.02639-5)=397.44 n/s
ce = 5 x 343.11 x (1.02639- 1) = 45.27 n/s
From this it is clear that the propagation ofthe compression wave is faster than the refer-
ence acoustic velocity, ao, and that the air particles move at a considerably slower rate. The
compression wave is moving rightward along the pipe at 397.4 m/s and, as it passes from
particle to particle, it propels each particle in turn in a rightward direction at 45.27 m/s. This is
deduced from the fact that the signs of the numerical values of ae and ce are the same.
The local particle Mach number, Me, is defined as the ratio of the particle velocity, ce, to
the local acoustic velocity, ae, where:
Me ce -
= 5 -1 (2.1.23)
164
From Eq. 2.1.12:
a. = aOX.
hence, ae=343.llxl.02639=352.16 mis
and the local particle Mach number, Me,
45.27
e 352.16
The mass rate ofgas flow, me caused by the passage ofthis point of the compression wave
in a pipe ofarea Ae is calculated from the thermodynamic equation of continuity as the multi-
plication of density, area, and particle velocity:
mie PeAece
From Eq. 2.1.17:
Pe = peX5 = 1.2049 x 1.026395 - 1.2049 x 1.1391 = 1.3725 kg/m3

The pipe area Ae is given by:
A
A~ td2
-- 3.14159x0.0252
= 0.000491 m
4 4
Therefore, because the arithmetic signs of the propagation and particle velocities are both
positive, the mass rate of flow is in the same direction as the wave propagation as:
= p'eAec = 1.3725 x 0.000491 x 45.27 = 0.0305 kg/s

The Expansion Wave
Second, consider the expansion wave of pressure pi:
pi = Pi x po = 0.8 x 101325= 81060 Pa

The pressure amplitude ratio, Xi, is calculated as:
Xi = 0.87 = 0.9686
165
Therefore, the propagation and particle velocities, as and ci, are found from:
ai = 343.11 x (6 x 0.9686 - 5) = 278.47 m/s
ci= 5 x 343.11 x (0.9686 - 1) = -53.87 m/s

From this it is clear that the propagation of the expansion wave is slower than the refer-
ence acoustic velocity, but the air particles move a little faster than the compression wave with
the same dp value. The expansion wave is moving rightward along the pipe at 278.47 m/s and,
as it passes from particle to particle, it propels each particle in turn in a leftward direction at
53.87 m/s. This is deduced from the fact that the numerical values of ai and ci are of opposite
sign.
The local particle Mach number, Mi, is defined as the ratio ofthe particle velocity to the
local acoustic velocity, ai, where:
M; = Ci
ai
From Eq. 2.1.12,
a1 = aoX,
hence ai =343.11x0.9686=332.3 m/s

and local particle Mach number, Mi,
-53.87
Mi = -0.1621
332.3
The mass rate of gas flow, fii, caused by the passage of this point of the expansion wave
in a pipe of area Ai is calculated by the continuity equation:
dih =p1Aici
From Eq. 2.1.17:
pi = piXi = 1.2049x0.96865 = 1.2049x0.8525= 1.0272 kim3
166
It is seen that the density is reduced in the more rarified expansion wave.
The pipe area is identical because the diameter is unchanged, i.e., Ai is 0.000491 m2. The
mass rate of flow is in the direction opposite to the wave propagation as:
in = p1A1c1 = 1.0272x 0.000491x (-53.87) = -0.0272 kgs
You should note that, in these compression and expansion waves with the same dp value,
the compression wave has the greater mass flow rate because its density is higher.
2.1.5 Distortion of the Wave Profile
It is clear from the foregoing that the value of propagation velocity is a function of the
wave pressure and wave temperature at any point on that pressure wave. It should also be
evident that, because all of the points on a wave are propagating at different velocities, the
wave must change shape in its passage along any duct. To illustrate this, the calculations
conducted in the previous section are displayed in Fig. 2.2. In Fig. 2.2(a) it can be seen that the
front and tail of the wave both travel at the reference acoustic velocity, ao, which is 53 m/s
slower than the peak wave velocity. In their travel along the pipe, the front and the tail will
keep station with each other in both time and distance. However, at the front ofthe wave, all of
the pressure points between it and the peak are traveling faster and will inevitably catch up
with it. Whether this will actually happen before some other event intrudes (for instance, the
wave front could reach the end of the pipe) will depend on the length of the pipe and the time
interval between the peak and the wave front. Nevertheless, there will always be the tendency
for the wave peak to get nearer the wave front and farther away from the wave tail. This is
known as "steep-fronting." The wave peak could, in theory, try to pass the wave front, which
is what happens to a water wave in the ocean when it "crests." In gas flow, "cresting" is
impossible and the reality is that a shock wave would be formed. This can be analyzed theo-
retically. Bannister [2.2] gives an excellent account of the mathematical solution for the
particle velocity and the propagation velocity, ash, of a shock wave of pressure ratio Psh propa-
gating into an undisturbed gas medium at reference pressure po and acoustic velocity ao0 The
theoretically derived expressions for propagation velocity, ash, and particle velocity, csh, of a
compression shock front are:
Psh =
Po
ash = ao 2fPsh +L2! = G67P5h +2G 17 (2.1.24)
167
Design and Simultion ofFour-Stroke Engines
(a) distortion of compression pressure wave profile
(b) distortion of expansion pressure wave profile

Fig. 2.2 Distortion of wave profile and possible shockformation.
csh = ah 0 a0(Psh - 1)
c G2Pa
ah +G1) (2.1.25)
These equations are derived in Appendix A2.2.

The situation for the expansion wave, shown in Fig. 2.2(b), is the reverse, in that the peak
is traveling 64.6 m/s slower than either the wave front or the wave tail. Thus, any shock forma-
tion taking place will be at the tail of an expansion wave, and any wave distortion will be
where the tail of the wave attempts to overrun the peak.
In the case of shock at the tail of the expansion wave, the above equations also apply, but
the "compression" shock is now at the tail of the wave and running into gas at acoustic state ai
and moving at particle velocity ci. Thus, the propagation velocity and particle velocity of the
168
shock front at the tail of the expansion wave, which has an undisturbed state at po and ao
behind it, are given by:
Psh = p
Pi
ash = ash relative to gasi + Ci
=aiv G67Ph +Gl7+ci (2.1.26)
= aoX G67Psh+GI7 +G5ao(Xi l)
Csh = Csh relative to gas i + ci

aiPh-lL +ci (2.1.27)
y G67Psh +GI7
aoXi(Psh -1) + G5aO(Xi -1)
y G67Psh +G17
For the compression wave illustrated in Fig. 2.2, the use of Eqs. 2.1.24 and 2.1.25 yields
finite amplitude propagation and particle velocities of 397.4 and 45.27 m/s, and for the shock
wave of the same amplitude values of 371.4 and 45.28 m/s, respectively The difference in
propagation velocity is some 7% less, but that for particle velocity is negligible.
For the expansion wave illustrated in Fig. 2.2, the use of Eqs. 2.1.26-and 2.1.26 yields
finite amplitude propagation and particle velocities of 278.5 and -53.8 m/s, and for the shock
wave of the same amplitude, values of 312.4 and 0.003 m/s, respectively The difference in
propagation velocity is some 12% greater, but that for particle velocity is considerable in that
the particle velocity at, or immediately behind, the shock is effectively zero.
2.1.6 The Properties of Gases
It will be observed that the propagation of pressure waves and the mass flow rate that they
induce in gases is dependent on the gas properties, particularly on the gas constant, R, and the
ratio of specific heats, y. The value of the gas constant, R, is dependent on the composition of
the gas. The value of the ratio of specific heats, y, is dependent on both gas composition and
temperature. It is essential to be able to index these properties at every stage of a simulation of
gas flow in engines. Much of this information can be found in many standard texts on thenno-
dynamics, but for reasons of clarity it is essential to repeat it here briefly, continuing the short
introduction given in Appendix Al.1.
169
The gas constant, R, of any gas can be found from the relationship relating the universal
gas constant R and the molecular weight, M, of the gas:
R=- R= (2.1.28)
M
The universal gas constant, R, has a value of 8314.4 J/kgmolK. The specific heats at
constant pressure and constant volume, Cp and Cv, are determined from their defined relation-
ship with respect to enthalpy, h, and internal energy, u:
C dh C= du (2.1.29)
dT dT
The ratio of specific heats, y, is found simply as:
CP
y =Cv (2.1.30)
Y-v
It can be seen that if the gases have internal energies and enthalpies that are nonlinear
functions of temperature, then neither Cp, Cv, nor y, are constants. If the gas is a mixture of
gases, then the properties of the individual gases must be assessed separately and then com-
bined to describe the behavior ofthe mixture.
To illustrate the procedure to determine the properties ofgas mixtures, let air be examined
as a simple example of a gas mixture with an assumed volumetric composition, u, of 21%
oxygen and 79% nitrogen, while ignoring the small, and relatively unimportant, 1% trace
concentration of argon. The molecular weights of oxygen and nitrogen are 31.999 and 28.013
respectively.
The average molecular weight of air, Mai., is then given by:
Mair = £(uvMps) = 0.21 x 31.999 + 0.79 x 28.013 = 28.85

The mass ratios, e, of oxygen and nitrogen in air are given by:
Ec
= 1Mq 0.21x31.999 0233
Mar 28.85
8'N2
N2MN2 0.79 x28.013 0.767
=076
Mar 28.85
170
The molal enthalpies, h, for gases are given as functions of temperature with respect to
molecular weight, where the K values are constants:
h = Ko +K1 T+K2T2 +K3T3 J/kgmol (2.1.31)
In this case, the molal intemal energy of the gas is related thermodynamically to the
enthalpy by:
u=h-RT (2.1.32)
Consequently, from Eq. 2.1.29, the molal specific heats are found by appropriate differen-
tiation of Eqs. 2.1.31 and 32:
Cp = Ki +22K2T+33K3T2 (2.1.33)
and UV =4 -R (2.1.34)
The molecular weights and the constants, K, for many common gases are listed in Table
2.1.1. These values are reasonably accurate for a temperature range of 300 to 3000 K. The
values of the molal specific heats, internal energies, and enthalpies ofthe individual gases can
be found at a particular temperature by using the values in Table 2.1.1.
Considering air as the example gas at a temperature of 20°C, or 293 K, the molal specific
heats of oxygen and nitrogen are found using Eqs. 2.1.33 and 2.1.34 as:
Oxygen, °2 C=31192 J/kgmolK Cv = 22877 J/kgmolK

Nitrogen, N2 Cp = 29043 J/kgmolK Cv = 20729 J/kgmolK
From a mass standpoint, these values are determined as follows:
Cp M C VCv
= (2.1.35)
M
Hence the mass related values are:
Oxygen, 02 Cp = 975 J/kgK Cv = 715 J/kgK

Nitrogen, N2 Cp = 1037 J/kgK Cv = 740 J/kgK
171
TABLE 2.1.1 PROPERTIES OF SOME COMMON GASES FOUND IN ENGINES.

GAS M l KO | K | K2|23
02 31.999 -9.3039E6 2.9672E4 2.6865 -2.1194E-4
N2 28.013 -8.503.3E6 2.7280E4 3.1543 -3.3052E-4
Co 28.011 -8.3141E6 2.7460E4 3.1722 -3.3416E-4
CO2 44.01 -1.3624E7 4.1018E4 7.2782 -8.0848E-4
H20 18.015 -8.9503E6 2.0781 E4 7.9577 -7.2719E-4
H2 2.016 -7.8613E6 2.621OE4 2.3541 -1.2113E-4
For the mixture of oxygen and nitrogen known as air, the properties are given generally as:
Rair = 7 (EgasRgas) CPair = z (egasCPgas)

(2.1.36)
CVair = X (egasCvgas) Yair =
CPair
CVair
Taking just one as a numeric example, the gas constant, R, which is not temperature
dependent, is found by:
I8314.4')
Rair = X(EgasRgas) = 0.233 31.999) + 0.767
83144'
= 288 J/kgK
3 28.0'11
and the specific heats for air at 293 K are found to be:
Cp = 1022 VkgK Cv = 734JkgK y = 1.393

It will be seen that the value ofthe ratio of specific heats, y, is not precisely 1.4 at standard
atmospheric conditions as stated earlier in Sec. 2.1.3. This is mostly because air contains
argon, which is not included in the above analysis. Because argon has a y value of 1.667, the
value deduced above is weighted downward arithmetically.
The most important point to note here is that these properties of air are a function of
temperature, so ifthe above analysis is repeated at 500 and 1000 K the following answers are
found for air:
T=500K Cp=1061 VkgK Cv=1773 YkgK y=1.373

T = 1000 K C = 1143 3kgK Cv = 855 J'kgK y = 1.337
172
Air can be found within an engine at these state conditions, thus is vital that any simula-
tion takes these changes of property into account because they have a profound influence on
the characteristics of unsteady gas flow.
Exhaust Gas
As a mixture of gases, exhaust gas clearly has quite a different composition compared to
air. This matter is discussed in much greater detail in Chapter 4, and was already introduced in
Chapter 1. However, consider now the simple and ideal case of stoichiometric combustion of
octane with air. The chemical equation, which has a mass-based air-fuel ratio, AFR, of 15, is as
follows:
2C8H18 +2102 + 79 N2J= 16CO2 +18H20+94.05N2
If the total moles is taken to be 128.05, then the volumetric concentrations ofthe exhaust
gas can be found by,
16 18 94.05
Uco2 __02=
= 128.05
__ = 0.125
.15 UH2O
UH = 128.05 = 0=41
__0.141 UN2 = 128.05
U20.734
This is precisely the same starting point as for the above analysis for air, so the procedure
is the same for the determination of all ofthe properties of exhaust gas that ensue from an ideal
stoichiometric combustion. A full discussion ofthe composition ofexhaust gas as a function of
air-fuel ratio is to be found in Chapter 4, Sec. 4.3.2, and an even more detailed debate in
Appendices A4.1 and A4.2, on the changes to that composition, at any fueling level, as a
function of temperature and pressure.
In reality, even at stoichiometric combustion there would be some carbon monoxide in
existence and minor traces of oxygen and hydrogen. Ifthe mixture were progressively richer
than stoichiometric, then the exhaust gas would contain greater amounts of CO and a trace of
H2, but would show little free oxygen. Ifthe mixture were progressively leaner than stoichio-
metric, the exhaust gas would contain lesser amounts of CO and no H2, but would show higher
concentrations of oxygen. The most important, perhaps obvious, issue is that the properties of
exhaust gas depend not only on temperature, but also on the combustion process that created
them. Tables 2.1.2 and 2.1.3 show the ratio of specific heats, y, and gas constant, R, ofexhaust
gas at various temperatures emanating from the combustion of octane at various air-fuel
ratios. An air-fuel ratio of 13 represents rich combustion; 15 is stoichiometric; and 17 is
approaching the normal lean limit ofgasoline burning. The composition ofthe exhaust gas is
shown in the Table 2.1.2 at a low temperature of 293 K. The influence of exhaust gas compo-
sition on the value of gas constant and the ratio of specific heats is quite evident. Although the
173
tabular values are quite typical of combustion products at these air-fuel ratios, naturally they
are approximate as they are affected by more than the air-fuel ratio. For example, the local
chemistry of the burning process and the chamber geometry, among many factors, will also
have a profound influence on the final composition of any exhaust gas. Table 2.1.3 shows the
composition of exhaust gas at higher temperatures. Compared with the data for exhaust gas at
293 K in Table 2.1.2, these same gaseous compositions show markedly different properties in
Table 2.1.3, when analyzed by the same theoretical approach.
From this it is evident that the properties of exhaust gas are quite different from air, and
while they are as temperature dependent as air, they are not influenced by air-fuel ratio, par-
ticularly with respect to the ratio of specific heats, as much as might be imagined. The gas
constant for rich mixture combustion of gasoline is some 3% higher than that at stoichiometric
and at lean mixture burning.
It is evident, however, that during any simulation of unsteady gas flow or ofthe thermody-
namic processes within engines, it is imperative for the accuracy of the simulation to use the
correct value of the gas properties at all locations within the engine.
2.2 Motion of Oppositely Moving Pressure Waves in a Pipe
In the previous section, you were asked to conduct an imaginary experiment with Fred,
who produced compression and expansion waves by exhaling or inhaling sharply, producing a
"boo" or an "u... uh," respectively. Once again, you are asked to conduct another experiment
so as to draw on your experience with sound waves to illustrate a principle-in this case the
behavior of oppositely moving pressure waves. In this second experiment, you and your friend
Fred are going to say "boo" to each other from some distance apart, and at the same time. Each
person's ears, being rather accurate pressure transducers, will record his own "boo" first, fol-
lowed a fraction of time later by the "boo" from the other party. Obviously, the "boo" from
each of you will have passed through the "boo" from the other and arrived at both Fred's ear
TABLE 2.1.2 PROPERTIES OF EXHAUST GAS AT LOW TEMPERATURE.

T=293 K % BY VOLUME
AFR % CO - °CO2 %H2O %O2 °hN2 R y
13 5.85 8.02 15.6 0.00 70.52 299.8 1.388
15 0.00 12.50 14.1 0.00 73.45 290.7 1.375
17 0.00 11.14 12.53 2.28 74.05 290.4 1.376
TABLE 2.1.3 PROPERTIES OF EXHAUST GAS AT ELEVATED TEMPERATURES.

T=500 K T=1000 K
AFR R y AFR R y
13 299.8 1.362 13 299.8 1.317
15 290.7 1.350 15 290.8 1.307
17 290.4 1.352 17 290.4 1.310
174
and your ear with no distortion caused by their passage through each other. If distortion does
take place, then the sensitive human ear is able to detect it. The point of meeting, when the
waves were passing through each other, is described as "superposition." This process occurs
continuously within the ducts of engines. The theoretical treatment below is for air. This sim-
plifies the presentation and will enhance your understanding of the basic theory; however,
extension of the theory to the generality of gas properties is straightforward.
2.2.1 Superposition of Opposiely Moving Waves
Fig. 2.3 illustrates two oppositely movig pressure waves in air in a pipe. They are shown
as compression waves, ABCD and EFGH, and are sketched as being square in profile, which
is physically impossible, but it makes the task of mathematical explanation somewhat easier.
In Fig. 2.3(a) the waves are about to meet. In Fig. 2.3(b) the process of superposition is taking
place for the front EF on wave top BC, and for the front CD on wave top FG. The result is the
creation of a superposition pressure, Ps' from the separate wave pressures, Pi and P2. Assume
that the reference acoustic velocity is ao. Assume also that the rightward direction is math-
ematically positive, and that the particle and the propagation velocity of any point on the wave
top BC will be cl and a,. From Eqs. 2.1.18 to 2.1.20:
c, = 5a0(XI -1) a, = a,(6X, - 5)
(a) two pressure waves approach each other in a duct
B
G
(b) two pressure waves partially superposed in a duct

Fig. 2.3 Superposition ofpressure waves in a pipe.
175
Similarly, with rightward regarded as the positive direction, the values for the wave top
FG will be:
C2 = -5a0(X2 -1) a2 = -ao(6X2-5)

From Eq. 2.1.14, the local acoustic velocities in the gas columns BE and DG during super-
position will be:
a, = a0XI a2 = aOX2
During superposition, the wave top F is now moving into a gas with a new reference
pressure level at Pl. The particle velocity of F relative to the gas in BE will be:
CF,lBE = - [(1 - 1J5aoXI [( -I]= -5ao(Xs -Xl)
The absolute particle velocity of F, c,, will be given by the sum of CFrelBE and cl, as
follows:
CS CFre1BE +Cl = -5aO(Xs-Xl)+5ao(XI -1) = 5ao(2X, -Xs -1)

Applying the same logic to wave top C proceeding into wave top DG gives another
expression for cs, as F and C are at precisely the same state conditions:
CS = CCrelDG +C2 = 5ao(Xs X2) 5a0(X2 -1) -5ao(2X2 Xs -1)

-
- = -
Equating the above two expressions for cS for the same wave top FC gives two important
equations as a conclusion, one for the pressure of superposition, Ps' and the other for the
particle velocity of superposition, cs:
Xs=Xi+X2-l (2.2.1)
or (Po () ( -o (2.2.2)
then PO Po (
then CS =5a%0(XI -l1)-5a0(X2 -1) = 5a0(XI - X2) (2.2.3)
176
or Cs =c1 +c2 (2.2.4)
Note that the expressions for superposition particle velocity reserves the need for a sign
convention, i.e., a declaration of a positive direction, whereas Eqs. 2.2.1 and 2.2.2 are inde-
pendent of direction. The more general expression for a gas with properties other than air is
easily seen from the above equations as:
Xs =X1 +X2-1 (2.2.5)
then IP I(P P - (2.2.6)
as Cs =c1 +c2 (2.2.7)
then cs = G5ao(X1 -1)-G5ao(X2 -1) = G5a0(X1 -X2) (2.2.8)
At any location within the pipes of an engine, the superposition process is the norm as
pressure waves continually pass to and fro. Further, if one places a pressure transducer in the
wall of a pipe, it is the superposition pressure-time history that is recorded, not the individual
pressures of the rightward and the leftward moving pressure waves. This makes it very diffi-
cult to interpret recorded pressure-time data in engine ducting. A simple example will make
the point. In Sec. 2.1.4 and in Fig. 2.2, an example is presented of two pressure waves, Pe and
pi, with pressure ratio values of 1.2 and 0.8, respectively. Suppose that these pressure waves
are in a pipe, but are the oppositely moving waves just discussed, and are in the position of
precise superposition. There is a pressure transducer at the point where the wave peaks coin-
cide and it records a superposition pressure, Ps. What will the value of Ps be and what is the
value of the superposition particle velocity, c.?
The values of Xe, Xi, ao, ce, and c; were 1.0264, 0.9686, 343.1, 45.3, and -53.9, respec-
tively, in terms of a positive direction for the transmission of each wave. In other words the
properties of the two waves, Pe and pi, are to be assigned to become those of waves 1 and 2,
respectively, merely to reduce the arithmetic clutter both within the text and in your mind.
If wave Pi is regarded as moving rightward and that is defined as the positive direction,
then the second wave, P2, is moving leftward in a negative direction. The application of Eqs.
2.2.1 and 2.2.2 shows:
Xs = X1 +X2 -1= 1.0264-0.9686-1=0.995

Ps =X7 = 0.965 P= Psp0 = 0.965 x 101325 = 97831 Pa
Cs = 45.3 + [-(-53.9)] = 99.2 mns
177
Thus, the pressure transducer in the wall of the pipe would show little of this process,
because the summation ofthe two waves reveals a trace that is virtually indistinguishable from
the atmospheric line, and exhibits nothing of the virtual doubling ofthe particle velocity.
The opposite effect takes place when two waves of the same type undergo a superposition
process. For example, if wave Pi meets another similar wave, pI, but going in the other direc-
tion and in the plane of the pressure transducer, then:
x = 1.0264+1.0264-1 = 1.0528 and Ps = Xs = 1.05287 = 1.434
Cs = 45.3 + -(+45.3 = 0 mn
The pressure transducer now shows a large compression wave with a pressure ratio of
1.434 and tells nothing of the zero particle velocity at the same spot and time.
This makes the interpretation of exhaust and intake pressure records within engine duct-
ing a most difficult business if it is based on experimentation alone. Not unnaturally, the
observing engineer will interpret a measured pressure trace exhibiting a large number ofpres-
sure oscillations as being evidence of lots of wave activity. This may well be so, but some of
these fluctuations will almost certainly be periods approaching zero particle velocity, while
yet other periods of the superposition trace exhibiting "calm" conditions could very well be
operating at particle velocities approaching the sonic value! This is clearly a most important
topic, and we will return to it in later sections ofthis chapter. It will also be a recurring theme
throughout the text.
2.2.2 Wave Propagation during Superposition
The propagation velocity of the two waves during the superposition process must also
take direction into account. The statement below is accurate for the superposition condition in
which the rightward direction is considered positive.
as = aoXs
The rightward superposition propagation velocity is given by the sum ofthe local acoustic
and particle velocities, as in Eq. 2.1.9:
as rightward = as + CS
= ao(XI +
X2 - 1) + G5ao(XI -
X2) (2.2.9)
= aO(G6XI - G4X2 -1)
178
as leftward = -as + Cs
= -aO(XI + X2 - 1) + G5a0(Xi X2) - (2.2.10)
= -aO(G6X2 - G4X1 - I)
During superposition in air of the two waves, Pe and pi, as presented above in Sec. 2.2.1,
where the values ofcs, a0, X1, X2, and X. were 99.2,343.1, 1.0264,0.9686, and 0.995, respec-
tively, this theory gives values ofpropagation velocity as:
a. = aOX =343.1xO.995=341.38 mds

asrightward = a. + Cs = 341.38 + 99.2 = 440.58 m/s
as leftward = -a. + Cs = -341.38 + 99.2 =-242.18 m/s

This could have been determined more formally using Eqs. 2.2.9 and 2.2.10 as:
asrightward = ao(G6XI - G4X2 - 1) = 343.38(6 x 1.0264 - 4 x 0.9686 - 1) = 440.58 m/s
as leftward = -a0(G6X2 - G4XI - 1) = -343.38(6 x 0.9686 - 4 x 1.0264 - 1) = -242.18 m/s

The original propagation velocities of waves 1 and 2, when they were travelling "undis-
turbed" in the duct into a gas at the reference acoustic state, were 397.44 m/s and 278.47 m/s.
It is thus clear from the above calculations that wave 1 has accelerated, and wave 2 has slowed
down, by some 10% during this particular example of a superposition process. This effect is
often referred to as "wave interference during superposition." In Sec. 6.4.5, you will be
referred back to this section for a "refresher course" when it becomes apparent that this effect
dominates the proceedings within a tuned exhaust pipe.
Consider some other basic examples, such as compression waves meeting each other, and
a similar encounter with expansion waves.
(i) Rightward wave Pe meets leftward wave Pe going in the opposite direction.
asrightward = a0(G6XI - G4X2 - 1) = 343.38(6 x 1.0264 - 4 x 1.0264 - 1) = 361.5 mn/s
asleftward = -a0(G6X2 - G4X1 - 1) = -343.38(6 x 1.0264 - 4 x 1.0264 - 1) = -361.5 m/s
179
Hence, the two compression waves meeting in superposition slow each other's propaga-
tion down by some 10%, and the particle velocity during superposition is zero.
(ii) Rightward wave pi meets leftward wave pi going in the opposite direction.
asrightward = a0(G6XI - G4X2 - 1) = 343.38(6 x 0.9686 - 4 x 0.9686 - 1) = 321.8 m/s
asleftward = -a0(G6X2 - G4X1 - 1) = -343.38(6 x 0.9686 - 4 x 0.9686 - 1) = -321.8 rn/s

Hence, the two expansion waves meeting in superposition accelerate each other's propa-
gation by some 15%, and the particle velocity during superposition is zero.
Because of the potential for arithmetic sign complexity, during computer calculations it is
imperative to rely on formal equation statements, such as Eqs. 2.2.9 and 2.2.10, to provide
computed values.
2.2.3 Mass Flow Rate during Wave Superposiion
Because the supexposition process accelerates some waves and decelerates others, the
mass flow rate must also be affected. It is therefore necessary to be able to calculate mass flow
rate at any position within a duct. The continuity equation provides the necessary information:
mass flow rate = densityx areax velocity= p,Ac,
where p = (2.2.11)
hence ifi = G5aopoA(X, + X2 - l)G5(XI - X2) (2.2.12)
In terms ofthe numerical example used in Sec. 2.2.2, the values of ao, po, X1, and X2 were
343.1, 1.2049, 1.0264, and 0.9686 respectively. The pipe area is that of the 25 mm diameter
duct, or 0.000491 m2. The gas in the pipe is air.
We can solve for the mass flow rate by using the previously known superposition value of
XS, which was 0.995, or the value for particle velocity, c3, which was 99.2 m/s, and determine
the superposition density, Ps, thus:
PS =PoXG5 = 1.2049x 0.9955 = 1.1751 kg'm3

Hence, the mass flow rate, mii, during superposition is given by:
rightward = 1.1751 x 0.000491 x 99.2 = 0.0572 kg/s
180
The sign could have been found by knowing that the superposition particle movement
was rightward and inserting cS as +99.2 and not -99.2. Alternatively, the formal equation,
Eq. 2.2.12, gives a numeric answer indicating the direction of mass or particle flow. This is
obtained by solving Eq. 2.2.12 with the lead term in any bracket, i.e., X1 as the value at which
wave motion is considered to be in a positive direction.
Hence, mass flow rate rightward, with the direction of wave 1 being called positive, is:
mrightward = GsaopoA(XI + X2 - I)G -X-X2)

= 5 x 343.11 x 1.2049 x (1.0264 + 0.9686 - 1) x (1.0264 - 0.9686)5
= +0.0572 kg/s
It will be observed, indeed it is imperative to satisfy the equation of continuity, that the
superposition mass flow rate is the sum ofthe mass flow rate induced by the individual waves.
The mass flow rates in the rightward direction of waves 1 and 2, computed earlier in Sec.
2.1.4, were 0.0305 and 0.0272 kg/s respectively.
2.2.4 Supersonic Particle Velocity during Wave Superposition
In typical engine configurations, it is rare for the magnitude of finite amplitude waves that
occur to provide a particle velocity approaching the sonic value. The Mach number, M, is
defined in Eq. 2.1.23 for a pressure amplitude ratio of X as:
M= c =G5a0(X -)= G5(X -) (2.2.12)

a a0X X
For this to approach unity then:
X= 5
G4
2-
3-y
andforair X=1.25
In air, as seen above, this would require a compression wave of pressure ratio P, where:
P= P =X'7 andinair P=1.257=4.768

Po
Even in high-performance racing engines, a pressure ratio for an exhaust pulse of greater
than 2.5 atmospheres is very unusual, thus sonic particle velocity emanating fom such a source
is not likely. A more realistic possibility is that a large exhaust pulse may encounter a strong
oppositely moving expansion wave in the exhaust system and the superposition particle veloc-
ity may approach, or attempt to exceed, unity.
181
Unsteady gas flow does not permit supersonic particle velocity. It is self-evident that the
gas particles cannot move faster than the pressure wave disturbance that is giving them the
signal to move. Because this is not possible according to gas dynamics, the theoretical treat-
ment supposes that a "weak shock" occurs and the particle velocity reverts to a subsonic value.
The basic theory can be found in any standard text [2.4] and the resulting relationships are
referred to as the Rankine-Hugoniot equations. The theoretical treatment is almost identical to
that given here for moving shocks, where the particle velocity behind the moving shock is also
subsonic (see Appendix A2.2).
Consider two oppositely moving pressure waves in a superposition situation. The indi-
vidual pressure waves are Pi and P2, and the gas properties are y and R, with a reference
temperature and pressure denoted by po and To. From Eqs. 2.2.5 to 2.2.8, the particle Mach
number, Mp, is found from:
Pressure amplitude ratio Xs= Xi +X2 -1 (2.2.14)
Acoustic velocity as = aoXs (2.2.15)
Particle velocity cs = G5ao(XI - X2) (2.2.16)
-L G5a0(X X2) (..7

-
Mach number (+ or!y) M =
(2.2.17)
Note that the modulus of the Mach number is acquired. This eliminates directionality in
any inquiry as to the magnitude of Mach number in absolute terms. If an inquiry reveals that
MS is greater than unity, then the individual waves P1 and P2 are modified by internal reflec-
tions to provide a shock to a gas flow at subsonic particle velocity. The superposition pressure
after the shock transition to subsonic particle flow at Mach number Msnew is labelled as Psnew.
The "new" pressure waves that travel onward after superposition is completed become
labeled as Plnew and P2new- The Rankine-Hugoniot equations describing this combined shock
and reflection process are:
Pressure p + y+1 (2.2.18)
M2+-2
M2 = -
Machnumber 2y M2-_ (2.2.19)
y-1
182
After the shock, the "new" pressure waves are related by:
Pressure Xsnew = XInew + X2nlw1 (2.2.20)
Pressure Pw= pX (2.2.21)
Mach number Msnew = G5ao(XI new - X2new) (2.2.22)

asnew aOXsnew
Pressure wave 1 = p 7Xn

xG (2.2.23)
Pressure wave 2 P2new = POx26ne (2.2.24)
Using the knowledge that the Mach number in Eq. 2.2.17 has exceeded unity, Eqs. 2.2.18
and 2.2.19 ofthe Rankine-Hugoniot set provide the basis for the simultaneous equations needed
to solve for the two unknown pressure waves Plnew and P2new through the connecting informa-
tion in Eqs. 2.2.20 to 2.2.22. For simplicity in presenting this theory, it is predicated that
PI > P2, i.e., that the sign of any particle velocity is positive. In any application ofthis theory,
this assumption must be borne in mind and the direction of the analysis adjusted accordingly.
The solution of the two simultaneous equations reveals, in terms of complex functions rV to
£4, which are composed of known pre-shock quantities:
M2+-2 -
+ 2 1 YM2L 1 ly-1
1=2y M2-_Iy+1 y +1 2
]r -
Xlnew _________-3r and X2new =-1

+ P4
- (2..25
then 2 (2.2.25)
The new values of particle velocity, Mach number, wave pressure, or other such param-
eters can be found by substitution into Eqs. 2.2.20 to 2.2.24.
Consider a simple numeric example of oppositely moving waves. The individual pressure
waves are Pi and P2 with strong pressure ratios of 2.3 and 0.5, and the gas properties are those
of air, where the specific heats ratio, y, is 1.4 and the gas constant, R, is 287 J/kgK. The
reference temperature and pressure are denoted by po and To, and are 101325 Pa and 293 K,
respectively. The conventional superposition computation as carried out previously in this
section would show that the superposition pressure ratio, P., is 1.2474, the superposition tem-
perature, Ts, is 39.1°(, and the particle velocity, cS, is 378.51 m/s. This translates into a Mach
183
number, Mp, during superposition of 1.0689, which is clearly just sonic. The application ofthe
above theory reveals that the Mach number, Msnew, after the weak shock is 0.937, and that the
ongoing pressure waves, Pinew and P2new, have modified pressure ratios of 2.2998 and 0.5956
respectively.
From this example, it is obvious that it takes waves of uncommonly large amplitude to
produce even a weak shock, and that the resulting modifications to the amplitude ofthe waves
are quite small. Nevertheless, this analysis must be included within any computational model-
ling of unsteady gas flow that has pretensions of accuracy. Sec. 6.4.5, and the discussion
related to Fig. 6.68, describe how this very effect delays the return of the expansion wave
reflection at the end of a tuned exhaust pipe.
In this section I have implicitly introduced the concept that the amplitude of pressure
waves can be modified by encountering some "opposition" to their perfect, i.e., isentropic,
progress along a duct. This also implicitly introduces the concept of reflections of pressure
waves, i.e., the taking of some ofthe energy away from a pressure wave and sending it in the
opposite direction. This theme is one that will appear in almost every facet ofthe discussions
that follow.
2.3 Friction Loss and Friction Heating during Pressure Wave Propagation
Particle flow in a pipe induces forces acting against the flow due to the viscous shear
forces generated in the boundary layer close to the pipe wall. Virtually any text on fluid
mechanics or gas dynamics will discuss the fundamental nature of this behavior in a compre-
hensive fashion [2.4]. The frictional effect produces a dual outcome: (1) the frictional force
results in a pressure loss to the wave opposite to the direction of particle motion and, (2) the
viscous shearing forces acting over the distance traveled by the particles with time means that
the work expended appears as internal heating ofthe local gas particles. The typical situation
is illustrated in Fig. 2.4, where two pressure waves, Pi and p2, meet in a superposition process.
waves superposed for time dt

Tw I- dx-
Pi %
Plf >0 '0
<0 P2
Ig
Lb
Fig. 2.4 Friction loss and heat transfer in a duct.
184
This makes the subsequent analysis more generally applicable. However, the following analy-
sis applies equally well to a pressure wave, Pl, traveling into undisturbed conditions, as it
remains only to nominate that the value of p2 is the same as that of the undisturbed state, po.
In the general analysis, pressure waves, Pi and P2, meet in a superposition process and due
to the distance dx traveled by the particles during a time dt engender a friction loss that gives
rise to internal heating, dQ; and a pressure loss, dpf. By definition, both of these effects
constitute a gain ofentropy, so the friction process is non-isentropic as far as the wave propa-
gation is concemed.
The superposition process produces all of the velocity, density, temperature, and mass
flow characteristics described in Sec. 2.2. However, the data regarding pressure loss and heat
generated, and more importantly the altered amplitudes ofpressure waves pi and p2 after the
friction process is completed, must be obtained from the theoretical analysis of friction pres-
sure loss and heating.
The shear stress, r, at the wall as a result of this process is given by:
Shear stress Cf Ps2 (2.3.1)
The friction factor, Cf, is usually in the range 0.003 to 0.008, depending on factors such as
fluid viscosity or pipe wall roughness. The direct assessment of the value of the friction factor
is discussed later in this section.
The force F exerted at the wall on the pressure wave by the wall shear stress in a pipe of
diameter d during the distance dx traveled by a gas particle during a time interval dt is
expressed as:
Distance traveled dx = csdt
Force F= drdTdx = ndtrcdt (2.3.2)
This force acts over the entire pipe flow area, A, and provides a loss ofpressure, dpf, for
the plane fronted wave that is inducing the particle motion. The pressure loss due to friction is
found by incorporating Eq. 2.3.1 into Eq. 2.3.2:
F - irdtc dt 4Tc dt 2Cfpsc3dt

Pesrlos dpf = S
Pressure loss A nd2 d d (2.3.3)
4
185
It will be observed that this equation contains a cubed term for the velocity, and because
there is a sign convention for direction, this results in a loss of pressure for compression waves
and a pressure rise for expansion waves, i.e., a loss of wave strength and a reduction of particle
velocity in either case.
Because this friction loss process is occurring during the superposition of waves of pres-
sure Pi and P2 as in Fig. 2.4, values such as superposition pressure amplitude ratio, Xs, density,
Ps, and particle velocity, cs, can be deduced from the equations given in Sec. 2.2. They are
repeated here:
XS =X1 +X2-1 Cs = G5ao(XI - X2) Ps = POXG5

S
The absolute superposition pressure, P., is given by:
Ps p-pXG7
= POxs
After the loss of friction pressure, the new superposition pressure, psf, and its associated
pressure amplitude ratio, X5f, will be, depending on whether it is a compression or expansion
wave,
Psf - Ps ± dpf (2.3.4)

Xsf = PPsf
The solution for the transmitted pressure waves, Plf and P2f, after the friction loss is
applied to both, is determined using the momentum and continuity equations for the flow
regime before and after the event, thus:
Continuity rhs = iflsf

Momentum PsA psfA misc5 nhsfcsf
- = -
which becomes A(ps - psf ) = sc. - ihisfcsf

The mass flow is found from:
iis = ps AcS = G5as (X1-X2)ApX
186
The transmitted pressure amplitude ratios, XIf and X2f, and superposition particle veloc-
ity, Csf, are related by:
Xsf = (Xlf + X2f -1) and Csf = G5ao(Xlf - X2f)

The momentum and continuity equations become two simultaneous equations for the two
unknown quantities, XIf and X2f, which are found by determining the particle velocity, csf:
Csf =C + Psf - Ps (2.3.5)

Pscs
where Xjf = Csf

212 + Xsf + G5a0 (2.3.6)
and X2f =l+Xsf -X1f (2.3.7)

Consequently the pressures of the ongoing pressure waves, PIf and P2f, after friction has
been taken into account, are determined by:
PIf =PoX1f and P2f = POX G7 (2.3.8)

Taking the data for the two pressure waves of amplitude 1.2 and 0.8 that have been used as
examples in previous sections ofthis chapter, consider these waves to be superposed in a pipe
of 25 mm diameter with the compression wave, P,, moving rightward. The reference conditions
are also as used before, that is, To is 20°C or 293 K and Po is 101,325 Pa. However, pressure
drop only occurs as a result of particle movement, so it is necessary to define a time interval
for the superposition process to occur. Let it be considered that the waves are superposed at
the pressure levels indicated for a period of 2r crankshaft in the duct of an engine running at
1000 rpm. This represents a time interval, dt, of:
0 60 0
dt5=-x-= s (2.3.9)
360 N 6N
2
or dt= 0.333 x10- 3 s
6 x 1000
187
where the particle movement, dx, is given by:
dx = csdt = 99.1x0.333x10-3 = 33x 10-3 m
From the above, the calculated numerical value for the superposition time element is
0.333 ms. Assuming a friction factor, Cf, of 0.004, the above equations in the section show
that the loss of superposition pressure occurs over a distance, dx, of 33 mm within the duct and
has a magnitude of 122 Pa. The pressure ratio of the rightward wave drops from 1.2 to 1.198,
and that ofthe leftward wave rises from 0.8 to 0.8023. The rightward propagation velocity of
the compression wave during superposition drops from 440.5 to 439.5 m/s, while that of the
leftward expansion wave rises from 242.3 to 243.4 m/s. The superposition particle velocity
drops from 99.1 m/s to 98.05 m/s.
It is evident that the pressure loss due to friction reduces the amplitude of compression
waves and slows them down. The opposite effect applies to an expansion wave: the pressure
loss due to friction raises its absolute pressure, i.e., weakens the wave, and thereby increases
its propagation velocity from a subsonic value toward sonic velocity.
Although the likelihood of a single traverse of a pressure wave in a duct of an engine is
remote, it should nevertheless be considered theoretically. By definition, friction opposes the
motion of a pressure wave and does so continuously. This means that a train of pressure waves
are sent offin the direction opposite to the propagation of the wave train and with a magnitude
that can be calculated from the above equations. Using the data above, but with the exception
that the single wave, PI, traveling rightward, i.e., in the positive direction according to the sign
convention, has a pressure ratio of 1.2. All other data remain the same and a fixed friction
factor of 0.004 is used. All ofthe above equations can be used, inserting a value Of P2 equal to
po, i.e., with a pressure ratio of 1.0. The results show that the ongoing wave pressure ratio, PIf,
is reduced to 1.1994 and the reflected wave, P2f, is 1.0004. If the calculation is repeated to
fnd the effect of friction on a single traverse of an expansion wave, i.e., by inserting a value of
0.8 for P1 and a value of 1.0 for P2, then Plf and Plf become 0.8006 and 0.9995, respectively.
In Sec. 2.19 the traverse of a single pressure wave in a duct is described as both theory and
experiment.
2.3.1 Friction Factor during Pressure Wave Propagaton
It is possible to predict the value of friction factor more closely by considering further
information available in the literature on experimental and theoretical fluid mechanics. The
thermal conductivity, Ck, and viscosity, j, of air, or any gas, are functions of absolute tem-
perature, T, and are required for the calculation of the shearing forces in air, from which fric-
tion factor can be assessed. The interconnection between friction factor and shear stress has
been set out in Eq. 2.3.1. The thermal conductivity and viscosity ofair can be found from data
tables and curves fitted to provide sufficiently accurate data for values of T from 300 to 2000 K
as:
Ck = 6.1944x10-3 +7.3814x10-5T-1.2491x10-8T2 W/nK (2.3.10)
188
=7.457x10 +4.1547x10-8T-7.4793x10-12T kgfms (2.3.11)

Although the data above for air are not the same as those for exhaust gas, the differences
are sufficiently small as to warrant describing exhaust gas by the same relationships used for
air for the purposes of determining Reynolds, Nusselt, and other dimensionless parameters
common in fluid mechanics.
The Reynolds number at any local point in space and time in a duct of diameter d at
superposition temperature T., density Ps, and particle velocity cs, is given by:
Re = psdcs (2.3.12)
IATS
Much experimental work in fluid mechanics has related friction factor to Reynolds num-
ber, and the expression ascribed to Blasius to describe filly turbulent flow is typical:
0.0791
Cf=
Re0.25
for Re24000 (2.3.13)
Almost all unsteady gas flow in engine ducting is turbulent, but where it is not, and it is
laminar or nearly laminar flow, it is simple and accurate to assign the friction factor as 0.01.
This threshold number can be easily deduced from the Blasius formula by inserting a Reynolds
number of 4000.
Using all of the data from the example cited in the previous section above, it was found
that the calculated superposition particle velocity is 99.1 m/s. The superposition density is
1.1752 kg/m3 and can be calculated using Eq. 2.2.11. The relationship for superposition tem-
perature is given by Eq. 2.1.11, therefore:
y-1
Ts = PS Y x2 (2.3.14)
Using the numerical data provided, the superposition temperature is 290.1 K or 17.1°C.
The viscosity ofair at 290.1 K is determined from Eq. 2.3.11 to be 1.888 x 10-5 kg/ms. Con-
sequently, the Reynolds number, Re, in the 25-mm diameter pipe, from Eq. 2.3.12, is 154,210;
it is clearly turbulent flow. Then, from Eq. 2.3.13, the friction factor, Cf, is calculated to be
0.00399, which is not far removed from the assumption used in the previous section and which
appears frequently for air in the literature!
189
The use of this theoretical approach for the determination of friction factor, Cf, allied to
the general theory regarding friction loss in the previous section, permits the accurate assess-
ment of the friction loss in pressure waves in unsteady flow.
It is important to stress that the action of friction on a pressure wave passing through the
pipe is a non-isentropic process. The manifestation of this is the heating of the local gas as
friction occurs and the continual decay of the pressure wave due to its application. The heat-
ing effect can be calculated because the friction force, F, is available from the combination of
Eqs. 2.3.1 and 2.3.2, and the work done by this force acting through distance dx appears as
heat in the gas element involved.
F = 7dcd[Cf Ps S c dt (2.3.15)
Thus the work, 8Wf, resulting in the heat generated, 8Qf, can be calculated by:
8Wf = Fdx = Fcsdt = C cd(2.3.16)

2
All of the relevant data for the numerical example used in this section are available to
insert into Eq. 2.3.16, from which it is calculated that the internal heating due to friction is
1.974 mJ. Although this value may appear miniscule, it should be remembered that this is a
continuous process occurring for a pressure wave during its excursion throughout a pipe, and
that this heating effect of 1.974 mJ takes place in a time frame of 0.333 ms. This represents a
heating rate of 5.93 W, putting the heating effect due to friction into a physical context that can
be more readily comprehended.
One issue that must be taken into account by those concerned with the computation of
wave motion is that all of the above equations use a length term within the calculation for
friction force with respect to the work done or heat generated by opposition to it. This length
term is quite correctly computed from the particle velocity cS and the time period dt for the
motion of those particles. However, should the computation method purport to represent a
group of gas particles within a pipe by the behavior of those particles at the wave point under
calculation, then the length term in the ensuing calculation for force F must be replaced by the
length occupied by the said group of particles. The subsequent calculation to compute the
work, SWf, i.e., the heat quantity aQf generated by friction, is the force due to friction for all of
the group of particles multiplied by the distance moved by any one of the particles in this
group, with this distance remaining as the "csdt' term. This is discussed at greater length in
Sec. 2.18.6.
190
2.3.2 Friction Loss during Pressure Wave Propagadon in Bends in Pipes

This factor is seldom considered of pressing significance in the simulation of engine duct-
ing because pressure waves travel around quite sharp kinks, bends, and radiused comers in
ducting with pressure loss not much greater than that normally associated with friction, as has
been discussed above. This is not a subject that has been researched to any great extent, as can
be seen from the referenced literature ofBlair et al. [2.20]. The basic mechanism of analysis is
to compute the pressure loss, dpb, in the segment of pipe length under analysis. The procedure
is similar to that for friction:
Pressure loss dpbdb=CbPscs

= 2 (2.3.17)
where the pressure loss coefficient, Cb, is principally a function of Reynolds number and the
deflection angle per unit length around the bend:
Cb = f(Re-,d-) (2.3.18)
dx
The extra pressure loss as a result of deflection, by an angle appropriate to the pipe seg-
ment length, dx, can then be added to the friction loss term in Eq. 2.3.3 and the analysis can
continue for the segment of pipe length, dx, under scrutiny. Sec. 2.14 contains a discussion of
pressure losses at branches in pipes, and Eq. 2.14.1 gives an almost identical relationship for
the pressure loss in deflecting flows around the corners of a pipe branch. All such pressure loss
equations relate that loss to a gain of entropy through a decrease in the kinetic energy of the
gas particles during superposition.
It is a subject deserving of painstaking measurement through research using the sophisti-
cated experimental QUB SP apparatus described in Sec. 2.19. The aim should be to derive
accurately the values of the pressure loss coefficient, Cb, and ultimately report them in the
literature; I am not aware of such information having been published already.
2.4 Heat Transfer during Pressure Wave Propagation
Heat can be transferred to or from the wall of the duct by the gas as the unsteady flow
process occurs. Although all three processes of conduction, convection, and radiation are
potentially involved, it is much more likely that convection heat transfer will be the predomi-
nant phenomenon in most cases. This is certainly true of induction systems, but some of you
will recall that exhaust manifolds glow red and ponder the potential errors of considering
convection heat transfer as the sole mechanism involved. There is no doubt that in such cir-
cumstances radiation heat transfer should be seriously considered for inclusion in the theoreti-
cal treatment. But this is not an easy topic and the potential error of including radiation could
actually be more serious than excluding it. As a consequence, only convection heat transfer
will be discussed.
191
The information needed to calculate the normal and relevant parameters for convection is
available from within the analysis of unsteady gas flow. The physical situation is illustrated in
Fig. 2.4. A superposition process is underway. The gas is at temperature Ts, particle velocity c5,
and density Ps.
In Sec. 2.3.1, the computation of the friction factor, Cf, and the Reynolds number, Re, was
described. From the Reynolds analogy of heat transfer with friction it is possible to calculate
the Nusselt number, Nu, thus:
Nu = CfRe (2.4.1)
2
The Nusselt number contains a direct relationship between the convection heat transfer
coefficient, Ch, the thermal conductivity ofthe gas, Ck, and the effective duct diameter, d. The
standard definitions for these parameters can be found in any conventional text on fluid
mechanics or heat transfer [2.4]. The definition for the Nusselt number is:
Nu = Chd (2.4.2)
Ck
From Eqs. 2.4.1 and 2.4.2, the convection heat transfer coefficient can be determined:
Ch = CkNu
d
= CkCfRe
2d
(2.4.3)
The relationship for the thermal conductivity ofair is given above in Eq. 2.3.10 as a fumc-
tion ofthe gas temperature, T. In any unsteady gas flow process, the time element, dt, and the
distance of exposure of the gas element to the wall, dx, is available from the computation or
from the input data. In such a case, knowing the pipe wall temperature, the heat transfer, SQh,
from the gas to the pipe wall can be assessed. The direction ofthis heat transfer is clear from
the ensuing theory:
SQh = idChdx(Tw - Ts )dt (2.4.4)

If we continue the numeric example with the same input data as given previously in Sec.
2.3, but with the added information that the pipe wall temperature is 100°C, then the output
data using the theory shown in this section give the magnitude and direction ofthe heat trans-
fer. Because the wall at 1001(C is hotter tan the gas at superposition temperature, Ts, at 17.10°C,
the heat tansfer direction is positive because it is added to the gas in the pipe. The value of
heat transferred is 23.44 mJ, which is considerably greater than the 1.974 mJ attributable to
friction. In this example, the Nusselt number, Nu, is calculated as 307.8 and the convection
heat transfer coefficient, Ch, as 326.9 W/m2K.
192
The total heating or cooling of a gas element undergoing an unsteady flow process is a
combination of the extemal heat transfer from convection and the internal heat transfer from
friction. This total heat transfer is defined as 8Qfh, and is obtained thus:
SQfh =6Qf +6Qh (2.4.5)

For the same numeric data, the value of dQfh is the sum of +1.974 mJ and +23.44 mJ, an
addition that would result in SQfh becoming +25.44 mJ.
2.5 Wave Reflections at Discontinuities in Gas Properties
As a pressure wave propagates within a duct, it is highly improbable that it will always
encounter ahead of it gas that is at precisely the same state conditions or has the same gas
properties as that through which it is currently traveling. In physical terms, many people have
experienced an echo when they have shouted in foggy surroundings on a cold, damp morning,
with the sun warming some sections of the local atmosphere more strongly than others. In
acoustic terms, this is precisely the situation that commonly exists in engine ducting. It is
particularly pronounced in the inlet ducting of a four-stroke engine that has experienced a
strong back-flow of hot exhaust gas at the onset of the intake process. Two-stroke engines
have an exhaust process that sends hot exhaust gas into the pipe during the blowdown phase
and then short-circuits a large quantity of much colder air into this same duct during the scav-
enge process. Racing four-stroke engines with long overlap valve periods and tuned exhaust
systems also induce some of the in-cylinder colder air into the hotter exhaust pipe. Pressure
waves propagating through such pipes encounter gas at varying temperatures, and reflections
ensue. Although this is referred to in the literature as a "temperature" discontinuity, this is
quite misleading. It is actually a discontinuity of state, so pressure wave reflection will take
place at such boundaries, even at constant temperature, if other gas properties such as the gas
constant or the density are variable across it.
The physical and thermodynamic situation is sketched in Fig. 2.5, where a pressure wave
p, is meeting pressure wave P2. Clearly, a superposition process takes place. In Section 2.2.1,
where the theory of superposition was set down, it was stated that once the superposition
process is completed the pressure waves, Pt and P2, proceed onward with unaltered ampli-
tudes. However, this superposition process is taking place with the arriving waves having
traversed through gas at differing thermodynamic states. Wave pi is coming from side "a"
where the gas has properties of ya and Ra, and reference conditions of density, POa' and tem-
perature, Toa. Wave P2 is coming from side "b" where the gas has properties of Yb and Rb, and
reference conditions of density, Pob, and temperature, Tob. As with most reflection processes,
the momentum equation is the equation that best describes a "bounce" behavior, and this proves
to be so in this case. Because the superposition process occurs, and the reflection is taking
place at the interface, the transmitted pressure waves assume amplitudes of PId and P2d. The
theory to describe this basically states that the laws of conservation of mass and momentum
must be upheld.
Continuity msidea = msideb (2.5.1)
193
discontinuity in gas properties

- -
P1 P2
side 'b' "al
gas on side 'a' gas on 1-4
0
Toa Tob +-a
0
Poa Pob
Ya Yb "al
Ra P2 Rb
Fig. 2.5 Wave reflection at a temperature discontinuity.
Momentum A(ps side a -Ps side b) = inside aCs side a - inside bcs side b (2.5.2)
It can be seen that this produces a relatively simple solution whereby the superposition
pressure and the superposition particle velocity are identical on either side of the thermody-
namic discontinuity.
Ps side a = Ps side b (2.5.3)
Cs side a = Cs side b (2.5.4)

The solution divides into two different cases, one simple and the other more complex,
depending on whether the gas composition is identical on both sides of the boundary.
(i) The Simple Case of Common Gas Composition
The following is the solution ofthe simple case where the gas is identical in composition
on both sides of the boundary, i.e., where ya and Ra are identical to Yb and Rb. Eq. 2.5.4 reduces
to:
G5aaoa(XI X2d) G5baob(Xld X2)

- = - (2.5.5)
Because the gas composition is common, this reduces to:
C : )bJ(XI -X2d) = (Xld - X2) (2.5.6)

5ObGS
(NOTE: Although the G5 terms are actually equal they are retained for completeness.)
194
Eq. 2.5.3 reduces to:
(XI + X2d_1)G7a = (Xld + X2 - 1)G7b (2.5.7)
In this simple case, the values of G7a and G7b are identical and are simply G7:
X1+X2d -1= Xld +X2-1 (2.5.8)

The solution becomes straightforward.
2X2 -Xrll- a(G5

X2d= Ob513) hence p - poX2d (2.5.9)
1+ a~G
aobG5b
Xld = X1 +X2d -X2 hence Pld = POXId (2.5.10)
(ii) The More Complex Case of Variable Gas Composition

If the gas composition is different across the boundary, the simplicity of reducing Eq.
2.5.6 to Eq. 2.5.7, and also Eq. 2.5.7 to Eq. 2.5.8, is no longer possible because Eq. 2.5.7
remains as a polynomial function. The method ofsolution is to eliminate one ofthe unknowns
from Eqs. 2.5.6 and 2.5.7, either Xld or X2d, and solve for the remaining unknown by the
Newton-Raphson method. The final step is as in Eqs. 2.5.9 and 2.5.10, but with the gas com-
position inserted appropriately to the side of the discontinuity:
= G7a (2.5.11)
Pd=PoX2d
p= G7b (2.5.12)
A numerical example will make the above theory easier to understand. Consider, as in Fig.
2.5, a pressure wave Pi arriving at a boundary where the reference temperature, ToI, on side
"a" is 200°C and T02 on side "b" is 100°C. The gas is colder on side "b" and is more dense. An
echo should ensue. The gas is air on both sides, i.e., y is 1.4 and R is 287 J/kgK. The situation
is undisturbed on side "b," i.e., pressure P2 is the same as the reference pressure po. The pres-
sure wave on side "a" has a pressure ratio, P1, of 1.3. Both the simple solution and the more
complex solution will give precisely the same answer. The transmitted pressure wave into side
"b" is stronger, with a pressure ratio, Pld' of 1.32, whereas the reflected pressure wave, the
echo, has a pressure ratio, P2d, of 1.016.
195
Ifthe above calculation is repeated with just one exception-the gas on side "b" is exhaust
gas with the appropriate properties, i.e., y is 1.36 and R is 300 J/kgK-then the simple solution
can be shown to be inaccurate. In this new case, the simple solution will predict that the
transmitted pressure ratio, Pld, is 1.328, whereas the reflected pressure wave, the echo, has a
pressure ratio, P2d, of 1.00077. The accurate solution, solving the equations and taking into
account the variable gas properties, will predict that the transmitted pressure ratio, Pld, is
1.3 14, whereas the reflected pressure wave, the echo, has a pressure ratio, P2d, of 1.011. The
difference in these answers is considerable, making the unilateral employment of a simple
solution, based on the assumption of a common gas composition throughout the ducting, inap-
propriate for use in engine calculations. In real engines, and therefore also in calculations with
pretensions of accuracy, a vanation of gas properties and composition throughout the ducting
is the nomn rather than the exception, and must be simulated as such.
2.6 Reflection of Pressure Waves
Oppositely moving pressure waves arise from many sources, but principally from reflec-
tions of pressure waves at the boundaries of the inlet or the exhaust duct. Such boundaries
include junctions at the open or closed end of a pipe, and also any change in area, gradual or
sudden, within a pipe. In the case ofsound reflections, an echo is a classic example of a closed-
ended reflection. All reflections are, by defmition, a superposition process, as the reflection
proceeds to move oppositely to the incident pressure wave causing it. Fig. 2.6 shows some of
the possibilities for reflections in a three-cylinder engine. Whether the engine is a four-stroke
or a two-stroke engine is ofno consequence, but the air inflow and the exhaust outflow at the
extremities are clearly marked. Expansion pressure waves are sent into the intake system by
the induction process and exhaust pressure waves are shuttled into the exhaust system in the
sequence of the engine firing order.
The potential for wave reflection in the ducting, and its location, is marked by a number. It
is obvious that the location accompanies a change of section of the ducting or a change in its
direction.
At point 1 is the intake bellmouth where the induction pressure wave is reflected at the
atmosphere. At point 2 is a plenum chamber where pressure wave oscillations are damped out
in amplitude. At point 3 is an air filter which provides a restriction to the flow and the possi-
bility of minor "echos" taking place. At point 4 is a throttle in the duct which, depending on the
throttle opening value, can pose either a major restriction or even none at all. At point 5 in the
intake manifold is a four-way branch where the pressure wave must divide and send reflec-
tions back from the change of section.
At point 6 in the exhaust manifold are three-way branches where the pressure wave must
also divide and send reflections back from the change of section. When a pressure wave
arrives at a three-way junction, the values transmitted into each ofthe other branches will be a
function of the areas of the pipes and all of the reflections present. When the induction
pressure wave arrives at the open pipe end at 1, a reflection will take place which will be
immediately superposed on the incident pressure pulse. In short, all reflection processes are
superposition processes-the fundamental reason why the theoretical text up to this point has
set down the wave motion within constant area ducting and the basics of the superposition
process in some detail.
196
15
I 1i
3
4 1:
5
7 r- -I
L-
9
Fig. 2.6 Wave reflection possibilities in the manifolds of an engine.
At points 7 and 8 are the valves or ports into the cylinder which, depending on the valve or
port opening schedule, behave as everything from a partially closed or open end to a cylinder
at varying pressures, to a perfect "echo" location when the valves or ports are all closed. At
point 9 are bends in the ducting where the pressure wave is reflected from the deflection
process in major or minor part depending on the severity of the radius of the bend. At point 10
in the exhaust ducting are sudden expansion and contractions in the pipe. At point 11 is a
tapered exhaust pipe which will act as a diffuser or a nozzle depending on the direction of the
particle flow; wave reflections ensue in both cases. At point 12 is a restriction in the form of a
catalyst, little different from an air filter except that chemical reactions are taking place at the
same time-an explanation that is easily written but whose underlying theoretical calculations
predicting the wave motion and the thermodynamics of the reaction are somewhat more com-
plex. At point 13 is a reentrant pipe to a chamber, which is a very common element in any
silencer design. At point 14 is an absorption silencer element, a length of perforated pipe
surrounded by packing, which acts as both diffuser and the trimmer of sharp peaks on exhaust
pulses. By definition, this element provides wave reflections. At point 15 is a plain-ended
exhaust pipe entering the atmosphere. Pressure wave reflections also take place here.
197
The above "tour" of the pressure wave routes in and out of an engine is far from a com-
plete description of the processes that take place. Nevertheless, it is meant to illustrate both
the complexity of the events and to postulate that, without a complete understanding of every
and all possibilities for wave reflection in and through an internal combustion engine, none
can seriously claim to be a designer of engines.
The resulting pressure-time histories are very complex and beyond the memory-tracking
capability of the human mind. A computer, however, is a methodical calculation tool that is
ideal for this pedantic exercise. You will therefore be introduced to the use of computers for
this purpose. Before that juncture, it is essential to comprehend the basic effect of each ofthese
reflection mechanisms, as the mathematics of their behavior must be programmed in order to
track the progress of all incident and reflected waves.
The sections that follow analyze virtually all ofthe above possibilities for wave reflection
due to changes in pipe or duct geometry, and analyze the reflection and transmission process
that takes place at each juncture.
2.6.1 Notation for Reflection and Transmission ofPressure Waves in Pipes
A wave arriving at a position where it can be reflected is called the incident wave. In the
paragraphs that follow, all incident pressure waves, whether they be compression or expansion
waves, will be designated by the subscript "i," i.e., pressure, pi; pressure ratio, Pi; pressure
amplitude ratio, Xi; particle velocity, ci; density, pi; acoustic velocity, ai; and propagation
velocity, as. All reflections will be designated by the subscript "r," i.e., pressure, pr; pressure
ratio, Pr; pressure amplitude ratio, Xr; particle velocity, cr; density, Pr; acoustic velocity, ar;
and propagation velocity, ar. All superposition characteristics will be designated by the sub-
script "s," i.e., pressure, Ps; pressure ratio, Ps; pressure amplitude ratio, Xs; particle velocity,
cs; density, ps; acoustic velocity, as; and propagation velocity, as.
Where a gas particle flow regime is taking place, flow from gas in a regime is always
subscripted with a "1" and flow to gas in a regime is subscripted with a "2." Thus the gas
properties of specific heats ratio and gas constant in the upstream regime are y, and R1, while
those in the downstream regime are y2 and R2.
The reference condition of pressure is noted as po; temperature as To; acoustic velocity as
ao; and density as po.
2.7 Reflection of a Pressure Wave at a Closed End in a Pipe
When a pressure wave arrives at the plane of a closed end in a pipe, a reflection takes
place. This is the classic echo situation, so it is no surprise to discover that the mathematics
dictate that the reflected pressure wave is an exact image ofthe incident wave, but traveling in
the opposite direction. The one certain fact available, physically speaking, is that the superpo-
sition particle velocity is zero in the plane of the closed end, as is shown in Fig. 2.7(a).
From Eq. 2.2.3:
Cs =c1 +Cr =0 (2.7.1)

or Cr =-Ci (2.7.2)
198
CSz
(a) closed end

<---- -Pr,
Pi I-
(b) open end outflov

)0- cs>O
--------
SS*SO
~ 4b-.'--1'11- l---|
(c) plain end inflow ~ I --r----
r61 Ps",,Po
Ps ..Po
..cs<O
Fig. 2.7 Wave reflection criteria at some typical pipe.
From Eq. 2.2.2, because cS is zero:
Xr =Xi (2.7.3)
hence Pr = Pi (2.7.4)
From the combination of Eqs. 2.7.1 and 2.7.3:
Xs =2X1-I (2.7.5)
therefore Ps = Po(2Xi - 1)G7 (2.7.6)
2.8 Reflection of a Pressure Wave at an Open End in a Pipe

Here the situation is slightly more complex, in that the fluid flow behavior is different for
compression and expansion waves. Compression waves are dealt with first. By definition,
such waves must produce gas particle outflow from the pipe into the atmosphere.
199
2.8.1 Reflection of a Compression Wave at an Open End in a Pipe

In this case, as illustrated in Fig. 2.7(b), the first logical assumption that can be made is
that the superposition pressure in the plane of the open end is the atmospheric pressure. The
reference pressure, po, in such a case is the atnospheric pressure. From Eq. 2.2.1, this gives:
Xs =X+Xr -1=1
hence Xr =2-Xi (2.8.1)
Therefore, the amplitude ofthe reflected pressure wave, pr, is:
Pr =po(2-Xi)G7 (2.8.2)
For compression waves, because Xi > 1 and Xr< 1, the reflection is an expansion wave.
From Eq. 2.2.6 for the more general case:
CS =G5aO(Xi-l)-G5ao(Xr -l)
=G5aO(Xi -1-2+Xi +1)
= 2G5aO(X; -1) (2.8.3)
= 2c;
This equation is applicable as long as the superposition particle velocity, cs, does not reach
the local sonic velocity, at which point the flow regime must be analyzed by further equations.
The sonic regime at an open pipe end is an unlikely event in most engines, but not all. An
analysis of this regime is described clearly by Bannister [2.2] and in Secs. 2.2.4 and 2.16 of
this text.
It can also be shown that, because cr equals c;, and because the equation is sign dependent,
the particle flow is in the same direction. This conclusion can also be reached from the earlier
conclusion that the reflection is an expansion wave that impels particles opposite to its direc-
tion of propagation. This means that an exhaust pulse arriving at an open end sends suction
reflections back toward the engine which will help to extract exhaust gas particles further
down the pipe and away from the engine cylinder-or out of it if the valve is open. Clearly,
this is a reflection that can be used by the designer to aid the scavenging, i.e., emptying ofthe
cylinder of its exhaust gas. The numerical data below emphasise this point.
For a basic understanding ofhow this theory is used to calculate the reflection of compres-
sion waves at the atmospheric end of a pipe, consider an example using the compression
pressure wave, Pe, previously used in Sec. 2.1.4.
You will recall that the wave, Peg is a compression wave of pressure ratio 1.2. In the
nomenclature of this section it becomes the incident pressure wave, pi, at the open end. This
200
pressure ratio is shown to have a pressure amplitude ratio, Xi, of 1.02639. Using Eq. 2.8.1, the
reflected pressure amplitude ratio, Xr, is given by:
Xr =2-X =2-1.02639=0.9736
or Pr = XG7 =0.97367 ==r.89

0.8293
That the reflection of a compression wave at the open end is a rarefaction wave is now
evident numerically.
2.8.2 Reflection of an Expansion Wave. at a Bellmouth Open End in a Pipe
This reflection process is connected with inflow, so it is necessary to consider the fluid
mechanics of the flow into a pipe. Inflow of air in an intake system, which is the normal place
to find expansion waves, is usually conducted through a bellmouth-ended pipe of the type
illustrated in Fig. 2.7. This form of pipe end will be discussed in the first instance.
The analysis of gas flow to and from a thermodynamic system, which may also be experi-
encing heat transfer and work transfer processes, is analyzed by the first law of thermodynam-
ics. The theoretical approach is to be found in any standard textbook on thermodynamics
[1.30, 2.4]. In general, this is expressed as:
A(heat transfer) + A(energyentering) = A(system energy) +

A(energyleaving) + A(worktransfer)
The First Law of Thermodynamics for an open system flow from the atmosphere to the
superposition station at the full pipe area in Fig. 2.7(d) is as follows:
c2 c2
Qsystem+AmO(ho + ) = dEsystem + Ams(hs + 2) + 8Wsystem (2.8.4)
If the flow at the instant in question can be presumed to be quasi-steady and steady-state
flow without heat transfer, and also to be isentropic, then SQ, SW, and dE are all zero. The
mass flow increments must be equal to satisfy the continuity equation. The difference in the
enthalpy terms can be evaluated as:
a2 _a2
h ho =Cp(Ts -TO)=- yR
y-I (Ts -TO)= y-lI
-
201
Although the particle velocity in the atmosphere, co, is virtually zero, Eq. 2.8.4 reduces
to:
c + G5ao2 2 2+
Gs (2.8.5)
Because the flow is isentropic, from Eq. 2.1.12:
as = aOXs (2.8.6)
Substituting Eq. 2.8.6 into Eq. 2.8.5 and now regarding co as zero:
c2 = G5a2(1- X2) (2.8.7)
From Eq. 2.2.8:
Cs = G5ao(Xi Xr) -
(2.8.8)
From Eq. 2.2.1:
Xs =Xi +Xr l1 (2.8.9)

Bringing these latter three equations together:
G5(Xi Xr)2 = l-(Xi + Xr -1)2

-
This becomes:
G6Xr - (2o4Xi + 2)Xr + (G6X? - 2Xi) = 0 (2.8.10)
202
This is an equation that is quadratic in Xr. Ultimately neglecting the negative sign in the
general solution to a quadratic equation, this yields:
(2G4Xi + 2) Xi4Xj +2) -4G6(G6Xi 2X) -

Xr =
-
(1+ G4Xi ji+ Xi (24 +2G6)+ Xi (G -G6) (2.8.11)

G6
For inflow at a bellmouth end when the incoming gas is air, where y equals 1.4, this
becomes,
X 1+4X+ 1+2oX-20X? (2.8.12)

r 6
This equation shows that the reflection of an expansion wave at a belimouth end of an
intake pipe will be a compression wave. Take as an example the expansion pressure wave pi of
Sec. 2.1.4, which by coincidence retains its subscript nomenclature because it represents an
incident wave, where that expansion wave has a pressure ratio, Pi, of0.8 and a pressure ampli-
tude ratio, Xi, of 0.9686. Substituting these numbers into Eq. 2.8.12 to determine the magni-
tude of the reflection of the wave at a beilmouth open end gives:
X 1 + 4x 0.9686+ Ii+ 20x 0.9686- 20x 0.96862 1.0238
Thus, rX7 = 1.02387 = 1.178

As predicted, the reflection is a compression wave and, if allowed by the designer to arrive
at the intake valve while the intake process is still in progress, it will push further air charge
into the cylinder. In the jargon of engine design, this effect is called "ramming." Ramming will
be discussed later in this chapter and even more thoroughly in Chapters 5 and 6.
203
2.8.3 Reflection of an Expansion Wave at a Plain Open End in a Pipe

The reflection of expansion waves at the end of plain pipes can be dealt with in a similar
theoretical manner. Bannister [2.2] describes the theory in some detail. It is presented here for
completeness, although the generality of this boundary condition is laid out in Sec. 2.16. You
can get a mental picture of the flow at this boundary by studying the results of a computational
fluid dynamics (CFD) calculation at this very discontinuity, presented in Fig. 3.3.
The First Law of Thermodynamics still applies, as seen in Eq. 2.8.5, and co is regarded as
effectively zero.
G0a =c +G5a2
G5a2=c2 +5aS (2.8.5)
However, as seen in Figs. 2.7(d) and 3.3, Eq. 2.8.6 cannot apply because the particle flow
has a distinct vena contracta within the pipe end with an associated turbulent vortex ring, and
cannot be regarded as an isentropic process. Thus:
as . aoXs
but an entropy gain raises the value of ao to aoo, thus:
as = aooXs (2.8.13)
The most accurate statement is that the superposition acoustic velocity, using the format of
Eq. 2.1.2, is given by:
a. = YS(2.8.14)
PS
Another gas dynamic relationship is required and, as is normal in a thermodynamic analy-

sis of non-isentropic flow, the momentum equation is employed. Consider the flow from the
atmosphere to the superposition plane of fillly developed flow downstream ofthe vena contracta.
Newton's Second Law of Motion can be expressed as:
Force = rate of change of momentum
where poA - psA = iiscs - ioco

or Po -Ps = Psc2 (2.8.15)
204
Using the wave summation equations, and ignoring the entropy gain encapsulated in the
aoo term so as to effect an approximate solution, the superposition pressure and particle veloc-
ity are related by:
Xs =Xi+Xrl-
CS = G5aO(Xi -
Xr)
=G7
PS = Y4
Elimination of unwanted terms does not provide a simple solution, but one that requires an
iterative approach by a Newton-Raphson method to arrive at the unknown value of reflected
pressure Xr from a known value ofthe incident pressure Xi at the plain open end. The solution
is expressed as a function of the unknown quantity Xr as follows:
f(Xr)=±[Xi+Xr j±X 1+ G (X X _-)G7 IXi =0 (2.8.16)
This is solved by the Newton-Raphson method, i.e., by differentiation of the above equa-
tion with respect to the unknown quantity, Xr, and iterating to find the answer in the classic
mathematical manner. Then the value of particle velocity, cr, can be derived from substitution
into the equation below once the unknown quantity, Xr, is determined.
Cs = Ci +Cr = Gao(Xi -Xr) (2.8.17)

This is a far from satisfactory solution to an apparently simple problem describing what is,
in reality, outflow from the atmosphere into a plain-ended pipe. That this is not simple at all
can be seen in Sec. 2.16.
2.8.4 The Inadequacy ofSimple Solutions
The above thermodynamics and gas dynamics, dealing with reflections ofpressure waves
at an open pipe end, probably appear sufficiently complex as to provide the definitive answer
to our problem-at least for the case of particle inflow into a pipe with a belhmouth end. I have
bad news, though: This is not correct. I have introduced these simple solutions in this section
to provide you with correct mental and numeric images ofwhat basically happens when pres-
sure waves arrive at an open pipe end. In reality, the inflow process at a plain open-ended pipe,
or a belimouth-ended pipe for that matter, is a singular manifestation of what is generally
called "cylinder to pipe outflow from an engine." In this case, the atmosphere is the very large
"cylinder" flowing gas into a pipe! This complex problem is treated in great detail later in Sec.
2.16. The present boundary conditions for inflow at a pipe end can be calculated by the same
205
theoretical approach as for pipe-to-cylinder flow, and the more complete thermodynamic and
gas dynamic statements appear in those pages. At this point, you may well feel that I have
wasted your time by asking you to study the simpler thermodynamics given above; not so, for
the more complex theory to come is no more than a logical extension of that which you have
been reading. Thus, when you get to Secs. 2.10, 2.12, 2.16, and 2.17, you will find comprehen-
sion that much easier. Even that is not the end ofthe matter, however, for you must wait until
Chapter 3, which details the discharge coefficients of plain-ended and bellmouth-ended pipes,
to learn how to solve these boundary conditions on a computer. However, I have good news,
too. Within the pages of this book are all the equations and the information that you need to
simulate the solution to these, and other, engine modeling problems. All you have to do is to
write the code.
2.9 An Introduction to Reflection of Pressure Waves at a Sudden Area Change
It is quite common to find a sudden area change within a pipe or duct attached to an
engine. In Fig. 2.6, sudden enlargements and contractions in pipe area are located at position
10. The basic difference-from a gas dynamics standpoint-between a sudden enlargement
and contraction in pipe area, such as at position 10, and a plenum or volume, such as at posi-
tion 2, is that the flow in the duct is considered to be one-dimensional, whereas in the plenum
or volume it is considered to be three-dimensional. A subsidiary definition is that the particle
velocity in a plenum or volume is considered to be so low as to be assigned as zero in any
thernodynamic analysis.
For reflections at area changes within the duct, treating the flow as one-dimensional gives
a change in amplitude of the transmitted pulse beyond the area change and also causes a wave
reflection from it. Such sudden area changes are sketched in Fig. 2.8, where it can be seen that
the pipe area contracts or expands at the junction. In each case, the incident wave at the sudden
area change is depicted as propagating rightward, with the pipe nomenclature being "1" for the
wave arrival pipe, with "i" signifying the incident pulse, "r" the reflected pulse, and "s" the
superposition condition. The notation "p" is the conventional symbolism for absolute pres-
sure. For example, at any instant, the incident pressure pulses at the junction are PiI and Pi2.
Depending on the areas A1 and A2, these incident pulses will give rise to reflected pulses, Pri
and Pr2.
In either expansion or contraction of the pipe area the particle flow is considered to be
proceeding from the upstream superposition station 1 to the downstream superposition station
2. Therefore, the properties and composition of the gas particles, which are considered to be
flowing, in any analysis based on quasi-steady flow are those of the gas at the upstream point.
In all of the analyses presented here, this nomenclature is maintained. Therefore, the various
functions of the gas properties are:
y=yl R=RI G5=G5, G7=G71,etc

It was Benson [2.4] who suggested a simple theoretical solution for such junctions. He
assumed that the superposition pressure at the plane of the junction was the same in both pipes
206
at the instant of superposition. This assumption is inherently one of an isentropic process.

Such a simple junction model will clearly have its limitations, but it is my experience that it is
remarkably effective in practice, particularly if the area ratio changes, Ar, are in the band,
6 <A <6
The area ratio is defined as:
Ar
rAl
=A2 (2.9.1)
From Benson,
PsI = Ps2 (2.9.2)
Consequently, from Eq. 2.2.1:
Xil +Xrl-i1 =Xi2+Xr2 1 (2.9.3)

From the continuity equation, equating the mass flow rate in an isentropic process on
either side of the junction where,
mass flow rate = density x area x particle velocity
pslAlcs, = Ps2A2Cs2 (2.9.4)
Using the theory of Eqs. 2.1.17 and 2.2.2, where the reference conditions are po, To and
Po, and with rightward decreed as positive particle flow, Eq. 2.9.4 becomes:
pOXG5AIG5aO(Xil- Xrl) -= pOXGXA2G5ao(Xi2 - Xr2) (2.9.5)
Because X., equals X52, this reduces to:
Al(Xil -Xrl) = -A2(Xi2 -

Xr2) (2.9.6)
207
Combining Eqs. 2.9.1, 2.9.3, and 2.9.6, and eliminating each ofthe unknowns in tun, i.e.,
Xri or Xr2:
Xrl = (1- Ar)Xii

1+Ar
+2Xi2Ar (2.9.7)
XXr2 +Ar)2..8
2Xil-(1Xi2(lAr)
To get a basic understanding of the results of employing Benson's simple "constant pres-
sure" criterion for the calculation of reflections of compression and expansion waves at sud-
den enlargements and contractions in pipe area, consider an example using the two pressure
waves, Pe and Pi, previously used in Sec. 2.1.4.
Wave Pe is a compression wave of pressure ratio 1.2 and wave pi is an expansion wave of
pressure ratio 0.8. Such pressure ratios have pressure amplitude ratios, X, of 1.02639 and
0.9686, respectively. Each of these waves in turn will be used as data for Xi, arriving in pipe 1
at a junction with pipe 2, where the area ratio will be either halved for a contraction or doubled
for an enlargement to the pipe area. In each case, the incident pressure amplitude ratio in pipe
2, Xi2, will be taken as unity, which means that the incident pressure wave in pipe 1 is facing
undisturbed conditions in pipe 2.
(a) An Enlargement, Ar = 2, for an Incident Compression Wave Where Pi, = 1.2 and
Xa = 1.02639
From Eqs. 2.9.7 and 2.9.8:
Xrl = 0.9912 Xr2 = 1.01759

Hence, the pressure ratios, Pr1 and Pr2, of the reflected waves are:
Prl = 0.940 Pr2 = 1.130

The sudden enlargement behaves like a slightly less effective "open end," as the com-
pletely open ended pipe in Sec. 2.8.1 gives a reflected wave with a pressure ratio of 0.8293
instead of0.940. The onward transmitted pressure wave into pipe 2 is also one of compression,
but it has a reduced pressure ratio of 1.13.
(b) An Enlargement, Ar = 2, for an Incident Expansion Wave Where Pi, = 0.8 and
Xi]= 0.9686
From Eqs. 2.9.7 and 2.9.8:
Xrl = 1.0105 Xr2 = 0.97908
208
Hence, the pressure ratios, Pr, and Pr2, ofthe reflected waves are:
Prl = 1.076 Pr2 = 0.862

As above, the sudden enlargement behaves as a slightly less effective "open end" because
a "perfect" bellmouth open end to a pipe in Sec. 2.8.2 was shown to produce a stronger
reflected wave with a pressure ratio of 1.178, instead of the weaker value of 1.076 found here.
The onward transmitted pressure wave in pipe 2 is one of expansion, but it has a diminished
pressure ratio of 0.862.
(c) A Contraction, Ar = 0.5, for an Incident Compression Wave where Pi, = 1.2 and
Xi]= 1.02639
From Eqs. 2.9.7 and 2.9.8:
Xrl = 1.0088 Xr2 = 1.0352

Hence, the pressure ratios, Pr, and Pr2, of the reflected waves are:
Prl = 1.063 Pr2 = 1.274

The sudden contraction behaves like a partially closed end, sending back a partial "echo"
of the incident pulse. The onward transmitted pressure wave is also one of compression, but
with an increased pressure ratio of 1.274.
(d) A Contraction, Ar O.S,for an Incident Expansion Wave where Pi, = 0.8 andXil = 0.9686
From Eqs. 2.9.7 and 2.9.8:
Xrl = 0.9895 Xr2 = 0.9582

Hence, the pressure ratios, Pr, and Pr2, of the reflected waves are:
Pri = 0.929 Pr2 = 0.741

As in (c), the sudden contraction behaves like a partially closed end, sending back a partial
"echo" of the incident pulse. The onward transmitted pressure wave is also one of expansion,
but it should be noted that it is now a stronger rarefaction wave with a pressure ratio of 0.741.
The theoretical presentation here, following that of Benson [2.4], is clearly too simple to
be completely accurate in all circumstances. The major objections to its use are that the
assumption of "constant pressure" at the discontinuity in pipe area cannot possibly be tenable
over all flow situations and that the underlying thermodynamic assumption is of isentropic
flow in all circumstances. It is, however, a very good guide as to the magnitude of pressure
209
Design and Sinulation ofFour-Stroke Engines
wave reflection and transmission. A more complete theoretical approach is examined in detail
in the following sections. A full discussion of the accuracy of Benson's simple assumption is
illustrated by numeric examples in Sec. 2.12.2.
Please be assured that you have not wasted your time by studying Benson's simple "con-
stant pressure" solution to a complex problem. Firstly, you now have a mental image of what
happens when pressure waves arrive at expansions and contractions in ducts. Secondly, the
more complex thermodynamic solution is one that is highly iterative on a computer. Because
computer time is a function ofthe number ofiterations required to gain the requisite accuracy
for the solution, and completion ofthat goal depends on avoiding the omnipresent potential for
the arithmetic "crashing" of such a calculation, the quality of the initial guess for the very first
iteration is extremely important. Benson's simple solution gives the perfect choice for that
initial guess.
2.10 Reflection of Pressure Waves at an Expansion in Pipe Area
This section contains the non-isentropic analysis of unsteady gas flow at an expansion in
pipe area. The sketch in Fig. 2.8(a) details the nomenclature for the flow regime, in precisely
the same manner as in Sec. 2.9. However, to analyze the flow completely, the further informa-
tion contained in the sketches of Figs. 2.9(a) arid 2.10(a) must also be considered.
Psi TS1 Ps2 TS2

PSI Csl Ps2 Cs2
Al.4Pr1
Pii R - Pr2 o A2
yl RI
Y2 R2
(a) sudden expansion in area in a pipe where cs5O
Psl Tsi Ps2 TS2

PsI CSi Ps2 CS2
A1
.4
--PrlI . P2 Pi2 iA2
Pil iR
yl Y2 R2
i RI
(b) sudden contraction in area in a pipe where cs>O
Fig. 2.8 Sudden contractions and expansions in area in a pipe.
210
UU------------ s PO P[I-
wu w
-------Tt------
DP - Po1=
0- TI-L
- ENTROPY - ENTROPY
(a) non-isentropic expansion (b) isentropic contraction
Fig. 2.9 Temperature entropy characteristicsfor simple expansions and contractions.
particle flow direction >

s2
Fig. 2.10 Particleflow in simple expansions and contractions.
In Fig. 2.10(a) the expanding flow is seen to leave turbulent vortices in the corners of the
larger section. That the streamlines of the flow give rise to particle flow separation implies a
gain of entropy from area section 1 to area section 2. This is summarized on the temperature-
entropy diagram in Fig. 2.9(a) where the gain of entropy for a diffusing flow rising from
pressure PsI to pressure Ps2 is clearly visible.
211
As usual, the analysis of flow in this quasi-steady and non-isentropic context uses, where
appropriate, the equations of continuity, the First Law of Thermodynamics, and the momen-
tum equation. The properties and composition of the gas particles are those of the gas at the
upstream point. Therefore, the vanrous functions of the gas properties are:
Y =YI R=R1 G5 =G51 G7 =G71,etc.

The continuity equation for mass flow in Eq. 2.9.5 is still generally applicable and
repeated here, although the entropy gain is reflected in the reference acoustic velocity and
density at position 2:
Inl-rn2=0 (2.10.1)
With rightward retained as the positive direction, this equation becomes:
pOXG5AIG5aO1(XjI -Xrl)+ po2XG5A2Gsao2(Xi2 - Xr2) = 0 (2.10.2)

The First Law of Thermodynamics was introduced in Sec. 2.8 for such flow situations.
The analysis required here follows similar lines of logic. The First Law of Thermodynamics
for flow from superposition station 1 to superposition station 2 can be expressed as:
2 2
hs +Cs'2 = hs2 + 2
or (cs2 +G5a2 ) (c 2 +G5a2)=o (2.10.3)
The momentum equation for flow from superposition station 1 to superposition station 2
is expressed as:
AlPs, +(A2 A1)ps, A2Ps2 +(Iislcsl liis2cs2)

- - - = 0
The logic for the middle term in the above equation is that the pressure PsI is convention-
ally presumed to act over the annulus area between the two ducts. The momentum equation,
also taking into account the information from the continuity equation regarding mass flow
equality, reduces to:
A2(PsI - Ps2) + as½(csi - C52) = 0 (2.10.4)
212
As with the simplified "constant pressure" solution according to Benson, presented in

Sec. 2.9, the unknown values will be the reflected pressure waves at the boundary, Pr, and Pr2'
and also the reference temperature at position 2, namely T02. There are three unlnowns,
necessitating three equations, namely Eqs. 2.10.2, 2.10.3, and 2.10.4. All other "unknown"
quantities can be computed from these values and from the "known" values. The known
values are the upstream and downstream pipe areas, A1 and A2, the reference state conditions
at the upstream point, the gas properties at superposition stations 1 and 2, and the incident
pressure waves, Pi, and Pi2*
It is recalled that:
G17 P
G17
Xii %Po - and Xi2 =
PoJ
The reference state conditions of density and acoustic velocity are:
Density Po P02= Po
Poi = RT01 (2.10.5)
RT02 (.05
Acoustic velocity ao0= yRT0 a02 = yRT02 (2.10.6)
The continuity equation, Eq. 2.10.2, reduces to:
pol(Xi1 + Xrl - l)G5A1G5aO1(Xij - Xrl)

(2.10.7)
+Po2(Xi2 + Xr2 - 1)G5A2GsaO2(Xi2 - Xr2) =0
The First Law of Thermodynamics, Eq. 2.10.3, reduces to:
[(G5aOi(Xil Xrl)) +G5aO (Xi1 + Xrl -
1)
[(G5aO2(Xi2 2 +G 2 +Xr2- l)10 (2.10.8)
213
The momentum equation, Eq. 2.10.4, reduces to:
POA2[(XiI + Xrl 1)G7 -(Xi2 + Xr2 1)G7]

+[POI(Xil + Xrl - 1)G5AjG5aoi(Xji - Xri)] (2.10.9)
x[G5aOi(Xij - Xrl) + G5ao2(Xi2 - Xr2)] = 0
The three equations cannot be reduced any further, as they are polynomial functions of all
three "unknown" variables. These functions can be solved by a standard iterative method for
such problems. I have found that the Newton-Raphson method for the solution of multiple
polynomial equations is stable, accurate, and rapid in execution. The arithmetic solution on a
computer is conducted by a Gaussian Elimination method.
As with all numerical methods, the computer time required is heavily dependent on the
number of iterations needed to acquire a solution of the requisite accuracy, in this case for an
error no greater than 0.01% for the solution of any of the variables. As forecast in Sec. 2.9, the
use of the Benson "constant pressure" criteria is invaluable in this regard because it consider-
ably reduces the number of iterations required. This is because the initial guess from the
Benson criteria is fairly close to the final answer. Numerical methods of this type are also
arithmetically "frail," if the user makes ill-advised initial guesses for the value of any of the
"unknowns." It is in this context that the use of the Benson "constant pressure" criteria is
particularly indispensable. Numeric examples are given in Sec. 2.12.2.
Isentropic Analysis for Diffusing Flow
Ifthe duct between section 1 and 2 in Fig. 2.10(a) has a gradual increase in area, such as in
a tapered pipe, rather than the sudden expansion illustrated, then it may be that the flow can be
considered to occur isentropically. The analysis above is for non-isentropic flow, as evidenced
by the theory showing that aO2 and T02 are not equal to ao, and ToI. To solve for isentropic
flow in this situation, these variables at section 2 now become constants:
Isentropic diffusion a02 = ao0 T02 =To,

In which case, as the number of variables in Eqs. 2.10.7 to 2.10.9 is reduced by one, so is
the number of relevant equations. Hence, because the momentum equation encapsulates the
non-isentropicity, it is abandoned, and the solution devolves to that for the continuity and
energy equations, Eqs. 2.10.7 and 2.10.8. This solution can be used for smoothly tapered pipes
in diffusing flow as long as the pipe taper is not too steep, and the particle velocity is not so
high that the particle flow separates from its attachment to the walls of the diff-user. See Sec.
2.15 for a more thorough discussion on this point.
214
2.10.1 Flow at Pipe Expansions Where Sonic Particle Velocity is Encountered

In the above analysis of unsteady gas flow at expansions in pipe area, the particle velocity
at section 1 will occasionally be found to reach, or even attempt to exceed, the local acoustic
velocity. In thermodynamic or gas dynamic terms this is not possible because the particles in
unsteady gas flow cannot move faster than the pressure wave signal that is impelling them.
The highest particle velocity that is permissible is the local acoustic velocity at station 1, i.e.,
the flow is pernitted to become choked. Therefore, during the mathematical solution of Eqs.
2.10.7, 2.01.8, and 2.10.9, the local Mach number at station 1 is monitored and retained at
unity if it is found to exceed it.
M = c =1-
asl
G5a01(Xij
a OXsj
-Xrl) G5(Xij -Xrl)
Xil +Xrl-1
(2.10.10)
This immediately simplifies the entire procedure as it gives a direct solution for one ofthe
unknowns:
if Ms, >11 Ms, = 1
hence Xrl = Ms, +Xi,(G5 -Ms,) - +G4Xi6 (2.10.11)

Ms,
+G5 G
The acquisition of all related data for pressure, density, particle velocity, and mass flow
rate at both superposition stations follows directly from the solution of the three polynomials
for Xrl, Xr2, and aO2, in the manner indicated in Sec. 2.9.
In many classic analyses ofchoked flow, a "critical pressure ratio" is determined for flow
from the upstream point to the throat where sonic flow is occurring. This concept cannot be
applied here, for it assumes zero particle velocity at the upstream point, which is clearly not
the case here for unsteady gas flow with this geometry.
2.11 Reflection of Pressure Waves at a Contraction in Pipe Area
This section contains the isentropic analysis of unsteady gas flow at a contraction in pipe
area. The sketch in Fig. 2.8(b) details the nomenclature for the flow regime, in precisely the
same manner as in Sec. 2.9. However, to analyze the flow completely, the further information
contained in the sketches of Figs. 2.9(b) and 2.1O(b) must also be considered.
In Fig. 2.10(b) the contracting flow is seen to flow smoothly from the larger section to the
smaller area section. The streamlines of the flow do not give rise to particle flow separation
and so the flow is considered to be isentropic. This is in line with conventional nozzle theory
as observed in many standard texts on thermodynamics. It is summarized in the temperature-
entropy diagram in Fig. 2.9(b) where there is no entropy gain for the flow falling from pres-
sure PsI to pressure Ps2*
215
As usual, the analysis of quasi-steady flow in this context uses, where appropriate, the
equations of continuity, the First Law of Thermodynamics, and the momentum equation. How-
ever, one less equation is required compared to the analysis for expanding or diffusing flow in
Sec. 2.10. This is because the value ofthe reference state is klown at superposition position 2
because the flow is isentropic:
To, = T02 or a0 =aO2 (2.11.1)

Because there is no entropy gain, the equation normally reserved for the analysis of non-
isentropic flow, the momentum equation, can be neglected in the ensuing analytic method.
The properties and composition of the gas particles are those of the gas at the upstream
point. Therefore, the various functions of the gas properties are:
y=yj R=R1 G5=Gr, G7=G7,,etc.

The continuity equation for mass flow in Eq. 2.9.5 is still generally applicable and is
repeated here:
il - m2 = 0 (2.11.2)
With rightward retained as the positive direction, this equation becomes:
po0XG5AjG5ao0(Xji - Xrl) + pO2XG5A2G5ao2(Xi2 - Xr2) = 0 (2.11.3)
or, XG5A1(Xil- Xrl) + XG5A2(XU2 - Xr2) = 0
2 2
hsi + Cs'2 = hs2 + 2
or (c21 +G a2 )(c2 +Ga2) = 0 (2.11.4)
As with the simplified "constant pressure" solution according to Benson, presented in

Sec. 2.9, the unknown values will be the reflected pressure waves at the boundary, PrI and p2.
There are two unknowns, necessitating two equations, namely Eqs. 2.11.3, and 2.11.4. All
other "unknown" quantities can be computed from these values and from the "known" values.
216
Chapter 2 - Gas Flew through Four-Stroke Engines
The known values are the upstream and downstram pipe areas Al andA2, the reference state
conditions at the upstream and downstream points, the gas properties at superposition stations
1 and 2, and the incident pressure waves, Pi, and pj2.
You will recall that:
G17 \G17
Pi
Xii = kPo J and Xi2 = Pi2J
~PO
The reference state conditions are:
Density PoI = P02 = RT (2.11.5)
Acoustic velocity ao0 = a02 = (2.11.6)
The continuity equation, Eq. 2.11.3, reduces to:
Po0(Xi + XrI -I )G5AjG5aO(Xjj - Xri)

+Po2(Xi2 + Xr2 - 0)G5A2G5ao2(Xi2 - Xr2) =0
or (Xui + Xrl 1)G5A(Xil- Xrl) + (Xi2 + Xr2 -1)G5A2(Xi2 Xr2) 0 (2.11.7)
The First Law of Thermodynamics, Eq. 2.11.4, reduces to:
[(G5aOl(Xii - Xrl)) + G5aOI(Xi + Xrl - 1)2]

[(G5aO2(Xi2 - Xr2)) + Gao2(Xi2 + Xr2 1)2] =
[G5(Xi Xrl)2+ (Xil + Xrl 1)2 -
(1
or
[G5(Xi2 Xr2)2 +(Xi2 +Xr2 -1) 2]=° (2.11.8)
217
These two equations cannot be reduced any further as they are polynomial functions ofthe
two variables. These functions can be solved by a standard iterative method for such prob-
lems. I have detennined that the Newton-Raphson method for the solution of multiple polyno-
mial equations is stable, accurate, and rapid in execution. The arithmetic solution on a
computer is conducted by a Gaussian Elimination method. Actually, this is not strictly neces-
sary, as a rather simpler solution can be effected as it devolves to two simultaneous equations
for the two unknowns and the corrector values for each of the unknowns.
Sec. 2.9 contains some comments about how the use of the Benson "constant pressure"
criteria, for the initial guesses for the unknowns to the solution, is indispensable. These com-
ments are still very appropriate. Numeric examples are given in Sec. 2.12.2.
rate at both superposition stations follows directly from the solution of the two polynomials
for Xri and Xr2.
2.11.1 Flow at Pipe Contractions where Sonic Particle Velocity is Encountered
In the above analysis of unsteady gas flow at contractions in pipe area, the particle veloc-
ity at section 2 can reach, and even attempt to exceed, the local acoustic velocity. In thermo-
dynamic or gas dynamic terms, this is not possible. The highest particle velocity that is
permissible is the local acoustic velocity at station 2, i.e., the flow is permitted to become
choked. Therefore, during the mathematical solution of Eqs. 2.11.7 and 2.11.8, the local Mach
number at station 2 is monitored and retained at unity if it is found to exceed it.
M c2 = G5aO2(Xi2- Xr2) G5(Xi2 -Xr2) (2.11.9)

as2 ao2Xs2 Xi2 +Xr2 -1
This immediately simplifies the entire procedure for this gives a direct solution for one of
the unknowns:
ifMS121 MSl=l
Xr2 = 2+ Xi2(G5 -MS2) = I+G4Xi2 (2.11.10)
In this instance of sonic particle flow at station 2, the entire solution can now be obtained
directly by substituting the value of Xr2 determined above into either Eq. 2.11.7 or 2.11.8 and
solving it by the standard Newton-Raphson method for the one remaining unknown, Xri
2.12 Reflection of Waves at a Restriction between Differing Pipe Areas
This section contains the non-isentropic analysis of unsteady gas flow at restrictions
between differing pipe areas. The sketch in Fig. 2.8 details much of the nomenclature for the
flow regime, but essential subsidiary information is contained in a more detailed sketch ofthe
218
Chapter 2 - Gas Flow through FourStroke Engines
geometry in Fig. 2.12. However, to analyze the flow completely, the further information con-
tained in the sketches of Figs. 2.11 and 2.12 must be considered completely. The geometry is
of two pipes of differing area, A1 and A2, which are butted together with an orifice of area At
sandwiched between them. This geometry gives rise to my jargon referrng to this discontinu-
ity as a "butted joint." This geometry is very common in engine ducting; for example, it could
be the throttle body of a carburetor with a venturi and a throttle plate. The details ofthis very
geometry are analyzed in some depth in Chapter 3.
It could also be simply a sharp edged, sudden contraction in pipe diameter, where A2 is
less than A1 and there is no actual orifice ofarea At at all. You may well feel that you have just
read all about that in Sec. 2.11, but the reality is that in this case the flow forms a vena contracta
with an effective area of throat At, which is less than A2. This requires a more complete theo-
retical treatment. In short, the theoretical analysis to be given here is the more accurate and
extended-and inherently more complex, version of that already presented in Sec. 2.11 for
sudden expansions and contractions in pipe area.
W ISENTROPIC LINE Psi

DI.
w
aT2
F-T2.
1's'NTROPY
Fig. 2.11 Temperature-entropy characteristicsfor a restricted area change.
pt, Tt
PsA, TS1 ct
Psl,) Csll Ps2, TS2
Al A2A
Fig. 2.12 Particleflow regimes at a restricted area change.
219
In Fig. 2.12, the expanding flow from the throat to the downstream superposition point 2
is seen to leave turbulent vortices in the corners of that section. That the streamlines of the
flow give rise to particle flow separation implies a gain of entropy from the throat to area
section 2. On the other hand, the flow from the superposition point 1 to the throat is contract-
ing and can be considered to be isentropic in the same fashion as the contractions discussed in
Sec. 2.11. This is su ized on the temperture-entropy diagram in Fig. 2.11, where the gain
of entropy for the flow rising from pressure Pt to pressure Ps2 is clearly visible. The isentropic
nature of the flow from pressure Psi to pressure Pt can also be observed as the vertical line on
Fig. 2.11.
The properties and composition of the gas particles are those of the gas at the upstream
point. Therefore, the various functions of the gas properties are:
Y=yl R=R1 G5=G51 G7=G71,etc.

As usual, the analysis of flow in this context uses, where appropriate, the equations of
continuity, the First Law of Thermodynamics, and the momentum equation.
Poi = Pot = Po Po
Density
RTO P02 RT02 (2.12.1)
=
Acoustic velocity ao0 = aOt =yRT0 aO2 = yRT (2.12.2)

The continuity equation for mass flow in previous sections is still generally applicable and
is repeated here, although the entropy gain is reflected in the reference acoustic velocity and
mil - rn2 =0
iii- t =0 (2.12.3)
With rightward retained as the positive direction, these equations become:
PoixG5A GaO (Xi, - Xrl) + p02X G5A2G5ao2(Xi2 - Xr2) = 0

p01xG5A1G5aO(XiI - Xrl) Pt[CcAt][Csct] = °
-
(2.12.4)
The above equation for the mass flow continuity for flow from the upstream station 1 to
the throat contains the coefficient of contraction on the flow area, Cc, and the coefficient of
220
velocity, C.. These are conventionally connected in fluid mechanics theory to a coefficient of
discharge, Cd, to give an effective throat area, At,ff, as follows:
Cd = CcCs and Atff = CdAt

The equation ofmass flow continuity, Eq. 2.12.4, now appears as:
P Si - Xrl) - CdptAtct = 0
The analysis required here follows similar lines of logic. The First Law ofThermodynamics
2 2
hs c = hs2 + 2
Si2
or (c21SGI
+G5a2
aSI-(c2 +G5a2 )=O
(C2 +Ga2)= (2.12.5)
The First Law of Thermodynamics for flow from superposition station 1 to the throat can
be expressed as:
2 2
hs1 + Csl = h + Ct
2 2
or Cp2(Ts -TO+ Si t =O (2.12.6)
The momentum equation for flow from the throat to superposition station 2 is expressed
as:
A2(pt Ps2) +rsi(ct - cs2) = 0

-
(2.12.7)
The unknown values will be the reflected pressure waves at the boundaries, Prl and Pr2,
the reference temperature at position 2, namely T02, and the pressure, Pt, and the velocity, ct, at
the throat. There are five unknowns, necessitating five equations, namely the two mass flow
equations in Eq. 2.12.4, the two First Law equations, Eq. 2.12.5 and Eq. 2.12.6, and the
momentum equation, Eq. 2.12.7. All other "'unknown" quantities can be computed from these
equations using the "known" values. The known values are the upstream and downstream pipe
areas, A1 and A2, the throat area, At, the reference state conditions at the upstream point, the
221
gas properties at superposition stations 1 and 2, and the incident pressure waves, PiI and Pi2. A
numeric value for the coefficient of discharge, Cd, is also required and further information on
this difficult and often controversial subject is provided in Chapter 3. For the Cd values at this
particular discontinuity of a "butted joint," consult Sec. 3.7.
You will recall that:
G17 G17 G17

Xii = Pii and Xi2= Pi2 andsetting Xt= PtJ
KPo i Po 'IPo)
Because there is isentropic flow from station 1 to the throat, the throat density and tem-
perature are given by:
Pt =poiXG5 and Tt = (aoiXt)

yR
The set of continuity equations in Eq. 2.12.4 reduce to:
Pol(Xi1 + Xrl - l)G5A1ao1(Xi - Xrl)

+pO2(Xi2 + Xr2 - l)G5A2ao2(Xi2 - Xr2) = 0 (2.12.8)
(Xi1 + Xrl-l)G5AG5aOj(Xi- Xrl)-XG5CdAtct =0 (2.12.9)
The First Law of Thermodynamics in Eq. 2.12.5 reduces to:
[(G5aOi(Xii Xrl)) +G5aOI(Xi, +Xrl- )

[(G5aO2(Xi2 Xr22 G 2 Xr2 l)]= (2.12.10)
G5 (aOI(Xil + Xrl )) (aOIXt)2} +(5aOI(Xl Xr ))2 -ct = (2.12.11)
222
The momentum equation in Eq. 2.12.7 reduces to:
pOA2[XtG7 (Xi2 + Xr2 - 1)G7]

+[POI(Xil + Xrl - l)G5AjG5aO(X;1 - Xrl))Ct + G5ao2(Xi2 - Xr2)I = 0 (2.12.12)
The five equations, Eqs. 2.12.8 to 2.12.12, cannot be reduced any further because they are
polynomial functions of all five "unknown" variables, Xrl, Xr2,, Xt, aO2, and ct. These func-
tions can be solved by a standard iterative method for such problems. I have found that the
Newton-Raphson method for the solution of multiple polynomial equations is stable, accurate,
and rapid in execution. The arithmetic solution on a computer is conducted by the Gaussian
Elimination method. Numeric examples are given in Sec. 2.12.2.
In this case, is not as easy to supply initial guesses that are sufficiently close to the final
answers to ease the arithmetic of the iteration process on the computer, as it was in previous
theoretical problems in this chapter. Although the Benson "constant pressure" assumption is
of great assistance in this matter, it is not the universal answer. Even with the assistance ofthis
assumption in arriving at a numerical solution, the theory must be programmed with great care
to avoid arithmetic instability during its execution.
2.12.1 Flow at Pipe Restrictions Where Sonic Paridcle Velocity is Encountered
In the above analysis of unsteady gas flow at restrictions in pipe area, the particle velocity
at the throat will quite commonly be found to reach, or even attempt to exceed, the local
acoustic velocity. In thermodynamic or gas dynamic terms, this is not possible. The highest
particle velocity that is permissible is the local acoustic velocity at the throat, i.e., the flow is
permitted to become choked. Therefore, during the mathematical solution of Eqs. 2.12.8 to
2.12.12, the local Mach number at the throat is monitored and retained at unity ifit is found to
exceed it.
As, Mt=-=1 Ithen ct=aoIXt (2.12.13)

aOiXt
This simplifies the entire procedure somewhat, for this gives a direct relationship between
two of the "unknowns," and so removes one of the previous equations from the solution. It is
probably easier from an arithmetic standpoint to eliminate the momentum equation, but from
a thermodynamic standpoint, it is certainly more accurate to retain it!
rate at both supexposition stations and at the throat follows directly from the solution of the
remaining four polynomials for Xrl, Xr2, Xt, a02, and ct.
223
2.12.2 Examples ofFlow at Pipe Expansions, Contractions and Restrictions

In Secs. 2.9 to 2.12 the theory of unsteady flow at pipe expansions, contractions, and
restrictions has been presented. In Sec. 2.9 the simple theory of a "constant pressure" assump-
tion by Benson was given whereas the more complete theory is in the subsequent sections. In
these various sections the point was made repeatedly that the simple theory is reasonably
accurate. Some numeric examples are given here which will illustrate that point and the extent
of the inaccuracy. The input and output data all use the nomenclature of the theory sections
and Figs. 2.8, 2.10 and 2.12. The input data for the diameters, d, are in mm units.
The input data include six different sets of numeric data. In Table 2.12.1 test data sets 1
and 2 illustrate an expansion of the flow, and test data sets 3 and 4 are for contractions. Test
data set 5 is for a restricted pipe at a contraction, whereas test data set 6 is for a restricted pipe
at an expansion in pipe area. The incident pulse for all tests has a pressure ratio of 1.2, the gas
is air and is undisturbed on side 2, and the reference temperature at side 1 is always 200C or
293 K. There are two output data sets: One in Table 2.12.2, where the Benson "constant pres-
sure" theory is used, and one in Table 2.12.3, where the more complex theory of Secs. 2.10,
2.11, and 2.12 is used as appropriate. The more complex theory gives information on the
pressure at the throat, Pt, and on the entropy gain exhibited by the reference temperature, T02.
The output shows the mass flow rate and the "error" when comparing the mass flow rate
computed by the "constant pressure" theory to that calculated by the more complex theory.
The column labeled "E or C" describes whether the flow encountered was at an expansion (E)
or a contraction (C) in pipe area.
For simple expansions, as in test data sets 1 and 2, the "constant pressure" theory works
remarkably well with an almost negligible error, i.e., less than 3% in terns ofmass flow rate.
Even ifthe expansion has a quite realistic coefficient of discharge of0.85 applied to it, then the
error on mass flow rate remains insignificant at 0.3% [3.18]. The magnitude of the reflected
pressure waves compare quite favorably in most circumstances.
For the simple, sudden contraction in test data sets 3 and 4, with an unrealistic coefficient
of discharge of 1.0, the mass flow rate error is 7.9% but changes little to a 9.1% error when a
potentially logical Cd value of 0.7 for a sharp-edged joint is applied [3.18]. The magnitude of
the reflected pressure waves resulting from the simple and the complex theories are not sig-
nificantly different.
TABLE 2.12.1 INPUT DATA TO THE CALCULATIONS

No. dl d2 d Cd Ar Pu PQ2
1 25 50 25 1.000 4.0 1.2 1.0
2 25 50 25 0.85 4.0 1.2 1.0
3 50 25 25 1.000 0.25 1.2 1.0
4 50 25 25 0.7 0.25 1.2 1.0
5 50 25 15 0.85 0.25 1.2 1.0
6 25 50 15 0.85 4.0 1.2 1.0
224
TABLE 2.12.2 OUTPUT DATA USING THE CONSTANT PRESSURE

THEORY OF SEC. 2.9
No. Pr1 Pn mg/s m error % E or C
1 0.8943 1.0763 45.15 -2.5 E
2 0.8943 1.0763 45.15 -0.3 E
3 1.1162 1.3357 52.68 7.90 C
4 1.1162 1.3357 52.68 9.10 C
5 1.1162 1.3357 52.68 42.3 C
6 0.8943 1.0763 45.15 45.7 E
TABLE 2.12.3 OUTPUT DATA FROM THE CALCULATIONS USING THE

MORE COMPLEX THEORY
No. Pr1 Pr2 mmg/s Pt Theory
1 0.8850 1.0780 46.34 1.065 SEC. 2.10
2. 0.8930 1.0770 45.31 1.061 SEC. 2.12
3 1.1230 1.3120 48.84 1.304 SEC. 2.11
4 1.1240 1.3080 48.25 1.279 SEC. 2.12
5 1.1460 1.2370 37.01 1.169 SEC. 2.12
6 0.9960 1.0540 30.98 1.030 SEC. 2.12
When any form of restriction is placed in the pipe, i.e., as in test data sets 5 and 6 for an
expansion and a contraction respectively, the "constant pressure" theory is capable of provid-
ing some relevant information for the magnitude of the reflected waves but not for the mass
flow rate. The error is similar for expansions than contractions, which is not the case for a
simple junction when restrictions between the pipes are not present. There are significant
entropy gains in tests 4 to 6.
Benson's "constant pressure theory" is seen to be reasonably accurate only for flow that
encounters simple, sudden expansions in the ducting.
2.13 An Introduction to Reflections of Pressure Waves at Branches in a Pipe
A simple theoretical treatment for this situation was also suggested by Benson [2.4] and in
precisely the same format as for the sudden area changes found in the previous section. A
sketch of a typical branch is shown in Fig. 2.13. The sign convention for all of the branch
theory presented here is that inward propagation of a pressure wave toward the branch will be
regarded as "positive." Benson [2.4] postulates that the superposition pressure at the junction,
at any instant of wave incidence and reflection, can be regarded as a constant. This is a straight-
forward extension of his thinking for the expansion and contractions given in Sec. 2.9.
The incident pressure waves are PiI' Pi2, and pi3, and the ensuing reflections are of pres-
sures Pri' Pr2, and Pr3. The superposition pressures are Psi, Ps2q and P3-
225
Xi3
a A3iy3nR3 1c3
(a) incidence and reflection
(b) at the junction faces
Fig. 2.13 Unsteadyflow at a three-way branch.
Benson s "Constant Pressure" Solution at a Three-Way Branch

A three-way branch is very similar to thejunction shown in Fig. 2.8, but with an extra pipe
joining in. Therefore, the theoretical solution involves the extension ofEqs. 2.9.3 and 2.9.4 to
deal with the superposition state and mass flow rate ofthe extra pipe 3 at thejunction. Benson's
criteria inherently assumes isentropic flow.
Psl = Ps2 = Ps3

or Xii +Xrl-= Xi2 +Xr2-1= Xi3 +Xr3 -1 (2.13.1)
The net mass flow rate at the junction is zero:
Al(Xi, - Xrl) + A2(Xi2 - Xr2) + A3(Xi3 - Xr3) = 0 (2.13.2)
226
There are three equations to solve for the three unknowns, Xrl, Xr2 and Xr3. It is presumed
that in the course of any computation, the values of all incident pressure waves from one
calculation time step to another are known quantities.
Where the total area, At, is defined as:
At =Al +A2+A3 (2.13.3)

The solution of the above simultaneous equations gives:
Xri = 2A2Xi2 + 2A3Xi3 + X l(Al A2 -A3)

-
(2.13.4)
At
X 2A1Xij +2A3Xi3 + Xi2(A2 A3 -A1)

- (2.13.5)
At
Xr3 2A1Xij+2A2Xi2 +Xi3X(A3 -A2 -Al) (2.13.6)

At
Perhaps not surprisingly, the branched pipe can act as either a contraction or an enlarge-
ment of area to the gas flow. In short, two pipes may be supplying one pipe, or one pipe
supplying the other two, respectively. Consider the two cases below, where the pressure waves
employed as examples are the familiar pulses that have been used so frequently already in the
text. For simplicity, all of the pipes are of equal area.
(a) Benson i Solution for a Compression Wave Arriving Down One of the Pipes
A compression wave comes down to the branch in pipe 1 through air, and all other condi-
tions in the other branches are "undisturbed." The compression wave has an incident pressure
ratio, Pi,, of 1.2, or the pressure amplitude ratio, Xi,, is 1.02639. The incident waves in the
other two pipes have pressure ratios and pressure amplitude ratios of unity.
The input to the calculation is:
Pil=1.2 Xi1=1.02639 Xi2=Xi3=1.0 as Pi2=Pi3=l.0

The result of the calculation using Eq. 2.13.3 is:
Xrl=0O9911 Xr2=Xr3=l .01759 hence Prl=0.940 and Pr2=Pr3=1.13
227
As far as pipe 1 is concerned, the outcome is exactly the same as that for a 2:1 area ratio
enlargement discussed in the previous section. At the branch, the incident wave divides evenly
between the other two pipes, transmitting a compression wave onward and reflecting a rarefac-
tion pulse into pipe 1. Because pipe 1 is supplying the other two pipes, the effect is an expan-
sion.
(b) Benson s Solution for Two Compression Waves Arriving Down Two of the Pipes
Compression waves of pressure ratio 1.2 arrive as incident pulses in pipes 1 and 2 leading
to the branch with pipe 3. Pipe 3 has undisturbed conditions, as its incident pressure ratio, Pi3,
is 1.0. Now the branch behaves as a 2:1 contraction as far as pipes 1 and 2 are concerned.
The input to the calculation is:
as Pi,= Pi2=1_2 Xi,= Xi2=l.02639 Xi3=l.0 as Pi3=l.0
The solution of Eq. 2.13.3 shows:
Xr1=Xr2=1.0088 Xr3=1.0352 hence Pr1=Pr2=1.0632 and Pr3=1.274

Pipes 1 and 2 are supplying the third pipe, hence the unsteady gas dynamic effect is that of
a contraction. The contraction effect is evidenced by the reflection of compression waves in
pipes 1 and 2 and the transmission of a compression wave onward into pipe 3. These numbers
are already familiar as computed data for pressure waves of identical amplitude at the 2:1
contraction discussed in the previous section.
When we have pipes with dissimilar areas and a mixture of compression and expansion
waves incident upon a branch, the situation becomes much more difficult for the human mind
to comprehend. At that point, the programming of the mathematics into a computer will leave
the designer's mind free to concentrate more on the relevance of the information calculated
and less on the arithmetic tedium of acquiring the data.
It is also obvious that the angles between the pipes at the branch must play some role in
determining the amplitudes of the transmitted and reflected waves. This makes it even more
obvious that a branch cannot transmit pressure waves without losses and entropy gains. This
subject was studied in recent times by Bingham [2.19], Blair [2.20], and Fleck et al. [2.68] at
QUB. Although the branch angles do have an influence on wave amplitudes, it is not as great
in some circumstances as might be imagined. So, for those who wish to ensure accuracy for
branched pipe calculations within an engine simulation, the theory in the following section is
essential.
2.14 The Complete Solution of Reflections of Pressure Waves at Pipe Branches
Both historically and theoretically, the next step forward was to attempt to solve the
momentum equation to cope with the non-isentropic realism that there are pressure losses for
real flows changing direction when moving around the sharp corners posed by the pipes at the
branch. Much has been written on this subject, and many of the references below provide a
commentary on the subject sustained over many years. Suffice it to say that the practical
228
approach adopted by Bingham [2.19], incorporating the use of a modified form ofthe momen-
tum equation to account for the pressure losses around the branch, is the basis of the method
used here. However, Bingham's solution was isentropic.
This same approach was also employed by McGinnity [2.39] in a non-isentropic analysis,
but for a single composition fluid only. His solution was further complicated by using a non-
homentropic Riemann variable method, which meant that, because gas properties were tied to
path lines, they were not as clearly defined as in the method used here. It would appear from
the literature [2.39] that the solution was reduced to the search for a single unknown quantity,
whereas I deduce below that there are actually five such unknown quantities for a complete
non-isentropic analysis of a three-way branch.
The merit of the Bingham method [2.19] is that it uses experimentally determined pres-
sure loss coefficients at the branches, an approach that is capable ofbeing enhanced further by
information from data banks, such as that published by Miller [2.38]. Further data are also
provided by Fleck et al. [2.68].
An alternative method, perhaps one that is more complete theoretically, is to resolve the
momentum of the flow at any branch into its horizontal and vertical components and equate
them both to zero. The disadvantage of this approach is that it does not include the fluid
mechanic loss component which the Bingham method incorporates so pragmatically and real-
istically.
Although the discussion here is devoted exclusively to three-way branches, the theoretical
process for the general case ofmultiple branches with n pipes, where n > 3, is almost identical
to that reported below. It will be seen that the basic approach is to identify the pipes that are the
suppliers, and the pipes that are supplied, at any junction. For an n-pipe junction, this is basi-
cally the only addition to the theory below, other than that the number of equations increases
by the number of extra junctions. Using the Newton-Raphson method for solution of multiple
polynomial equations, and handling the matrix arithmetic by the Gaussian Elimination method,
the additional computational complexity is negligible.
The sign convention for particle flow is declared as "positive" toward the branch, and Fig.
2.13 inherently stipulates this convention. The pressure loss criteria for flow from one branch
to another is set out by Bingham as,
Pressure loss AP=CLPscS2

Ap= (2.14.1)
where the loss coefficient, CL, is given by the inter-branch angle, 0, in units of degrees:
if 0 < 167° then CL = 1.6- 1.6- else CL = 0 (2.14.2)

167
In any branch there are supplier pipes and supplied pipes. There are two possibilities in
this regard and these lead to two assumptions for their solution. The more fundamental
assumption in much of the theory is that the gas within the pipes is a mixture of two gases; in
an engine context these are logically exhaust gas and air. Obviously, the theory is capable of
229
being extended to mixtures of many gases, as indeed air and exhaust gas actually are. In addi-
tion, the theory is capable of being extended relatively easily to branches with any number of
pipes at the junction. For greater ease of understanding, the theory set out below details a
three-way branch for greater ease of understanding.
(a) One Supplier Pipe at a Three-Way Branch
Here there is one supply pipe and two pipes that are being supplied. In this case, the
required solution is for the reflected wave amplitudes, Xr, in all three pipes, and for the refer-
ence acoustic velocities, a0, for the gas going toward the two supplied pipes. (NOTE: The
word "toward" is used here precisely.) The consequence ofthis is that the mathematical solu-
tion has five unknown values needing five equations. It is possible to reduce the number of
unknowns by one, ifwe assume that the reference acoustic state toward the two supplied pipes
is common. A negligible loss of accuracy accompanies this assumption. Using the notation of
Fig. 2.13, it is implied that pipe 1 is supplying pipes 2 and 3; this notation will be used here
only to "particularize" the solution, so as to aid understanding of the analysis. Because the
pressure in the face ofpipes 2 and 3 will normally be nearly the same, the difference between
T02 and T03 should be small and, because the reference acoustic velocity is related to the
square root of these numbers, the error is potentially even smaller. Irrespective of this assump-
tion, it then follows absolutely that the basic properties ofthe gas entering the supplied pipes
are those of the supplier pipe.
(b) Two Supplier Pipes at a Three- Way Branch
Here there is one supplied pipe and two pipes that are suppliers. In this case, the required
solution is for the reflected wave amplitudes, Xr, in all three pipes, and for the reference acous-
tic velocity, ao, in the supplied pipe. This means that there are four unknown values needing
four equations. This is only possible by making the assumption that the superposition pres-
sures at the faces of the two supplier pipes are equal; this is the same assumption used by
Bingham and Blair [2.19] and McGinnity [2.39]. It then follows that the gas entering the
supplied pipe has the properties of a mass flow related mixture of the gases in the supplier
pipes. Using the notation of Fig. 2.13, it is implied that pipes 1 and 2 are supplying pipe 3, so
the subscripts of the equations are juxtaposed if the supplied/supplier pipe scenario changes.
The subscript notation "e3" should be noted carefully, for this details the quantity and
quality of the gas entering, i.e., going "toward," the mesh space beyond the pipe 3 entrance,
whereas the resulting change of all of the gas properties within that mesh space is handled by
the unsteady gas dynamic method of Sec. 2.18.9.
Continuity me3 = lhl + th2 (2.14.3)
Purity e= mrII+rn2H2
m43 (2.14.4)
230
Gas constant Re3 = + 2 (2.14.5)

me3
Specific heats ratio Ye3 = - lCP1 + xh2CP2 (2.14.6)

C Ve3 mIlCVI + ii2CV2
The final analysis then relies on incorporating the equations resulting from all of these
previous considerations regarding pressure losses, together with the continuity equation and
the First Law ofThermodynamics. The notation ofFig. 2.13 applies, together with either Fig.
2.14(a) for two supplier pipes, or Fig. 2.14(b) for one supplier pipe.
ISENTROPIC LINE Psi
w
I1 Ps2
3
w
CL T
(a) one supplier pipe to the branch

ISENTROPIC LINE Psi
/P Ps2
w
cc
Q
IIJ
cc
w
0-
w
(a) two supplier pipes to the branch
Fig. 2.14 The temperature-entropy characteristics.
231
Design and Sinulation ofFour-Stroke Engines
Density, Particle Velocity, and Mass Flow Rate at a Three-Way Branch

(a) For One Supplier Pipe
For one supplier pipe, the following are the relationships for the density, particle velocity,
and mass flow rate that apply to the supplier pipe and the two pipes that are being supplied:
Psi = Pol Xii + Xrl 1)G51 csl = G5iaol(Xil - Xrl) ihl = PsIAIcsI (2.14.7)
Ps2 = POe2 Xi2 + Xr2 -1) cs2 = G5laOe2(Xi2 - Xr2) m2 = Ps2A2cs2 (2.14.8)
Ps3 = POe3 Xi3+Xr3 -1)G51 Cs3 = G5iaOr3(Xi3 - Xr3) r3 = Ps3A3cs3 (2.14.9)
(b) For Two Supplier Pipes

For two supplier pipes, where the assumption is that the superposition pressure in the
faces of pipes 1 and 2 are equal, the equations for pipes 2 and 3 become:
Ps2 = P02(Xi2 + Xr2 -1)G52 cs2 = G52aO2(Xi2 - Xr2) mn2 = Ps2A2Cs2 (2.14.10)
Ps3 = POe3(Xi3 + Xr3)-1 css= G5lac3(Xi3 - Xr3) mP3

= Ps3A (2.14.11)
The First Law of Thermodynamics is expressed as,
ilhOhI + lii2hO2 + ii3ho3 = 0 (2.14.12)

where the local work, heat transfer and system state changes are logically ignored, and the ho
term is the stagnation specific enthalpy. This single line expression is the one to be used if the
assumption is made that T02 equals T03 for a single supplier pipe situation; it is also strictly
correct for a two supplier pipes model. To solve this equation without this assumption, it must
be split into two and analyzed specifically for the separate flows from pipe 1 to 2 and from
pipe 1 to 3. The reference densities in question result from this step, as shown below. The
normal isentropic expressions for the reference densities are expressed as:
Poi RTR
=
PoP0=
1ol
Po
P02- R2T02 P03=P30
PT 03TR3 (2.14.13)
232
However, should gas be entering pipes 1 or 2 as a result of a non-isentropic process, then

these expressions become, for pipes 2 and 3,
POe2= Po P3 =-Po (2.14.14)

Re2To2 Re3To3
Stagnation Enthalpy
Stagnation enthalpy is generally defined as the addition of enthalpy and kinetic energy as
follows:
ho =h+ 2 =CpT+ 22
(a) Stagnation Enthalpyfor One Supplier Pipe

Where one pipe (in this example, pipe 1) is the supply pipe, the stagnation enthalpies are:
G12 2 +2 G52a2X2 2 G,32 2 +2

=o Ga01X~2 +csl h02 e~X2
2
+c s2 Gh
h03- a3Xs3
2
s3 (2. 14.15)
(b) Stagnation Enthalpyfor Two Supplier Pipes

As with continuity, when there are two supplier pipes, only the expression for stagnation
enthalpy for pipe 2 is altered; the other expressions in Eq. 2.14.15 remain the same:
G2aX 2
ho2 = G52a02Xs2 +Ci2 +2 (2.14.16)
2
Pressure Losses at the Three-Way Branch
The pressure loss equations first presented in Eq. 2.14.2 devolve to:
(a) For One Supplier Pipe
For one supplier pipe, the flow is from pipe 1 to pipes 2 and 3,
CL12 = 1.6- 1.6012 CLI3 = 1.6- 1-6013 (2.14.17)

167 167
SIxW - W2) CLUP2p2cs2 P(SI - 73) -CL13pS3cs3 (2.14.18)
233
(b) For Two Supplier Pipes

Here there are two supplier pipes, i.e., pipes 1 and 2 are supplying pipe 3,
CL12 = 0 CL13= 1.6 1.6013

167
(2.14.19)
XS1-X =0 PO(XSlS-IXs3)-=CL13PS3CS3 (2.14.20)
In the analysis for the First Law of Thermodynamics, and the pressure loss terms, it should
be noted that the superposition relationships for pressure amplitude ratio and particle velocity,
X. and cs, for each pipe are written in full in the continuity equation to reveal the unknown
variables, Xr2, Xr2, and Xr3.
These functions can be solved by a standard iterative method for such problems. I find the
Newton-Raphson method for the solution of multiple polynomial equations to be stable, accu-
rate, and rapid in execution. The arithmetic solution on the computer is conducted by a Gaussian
Elimination method.
It should be noted that the "unknowns" are the three reflected wave pressures and T02
and/or T03, depending on the thermodynamic assumptions discussed earlier. In practice, it has
been found that Benson's "constant pressure" criterion provides excellent initial guesses for
the unknown variables. This permits a final numerical solution to be obtained in a mere two or
three iterations with a maximum error of only 0.05% of any one of the values.
This more sophisticated branched pipe boundary condition can be incorporated into an
unsteady gas dynamic code for implementation on a digital computer. Papers describing such
work have already been presented [2.41, 2.59, 2.68].
2.14.1 The Accuracy of Simple and More Complex Branched Pipe Theories
The assertion is made above that the "constant pressure" theory of Benson is reasonably
accurate. The following example puts numbers to that statement. Consider a branch consisting
of three pipes numbered 1 to 3, where the initial reference temperature of the gas is 200C and
the gas in each pipe is air. The angle 012 between pipe 1 and pipe 2 is 300 and the angle 013
between pipe 1 and pipe 3 is 1800, i.e., pipe 3 is lying straight through from pipe 1. The input
data are shown in Table 2.14.1. The pipes are all of equal diameter, d, at 25 mm in tests 1, 3,
and 4 in Table 2.14.1. In test 2, the diameter of pipe 3, d3 is 35 mm. In tests 1 and 2 the incident
pulse in pipe 1 has a pressure ratio, Pi, of 1.4 and the other pipes have undisturbed wave
conditions. In test 3 and 4, the incident pulse in pipe 1 has a pressure ratio, Pi, of 1.4 and the
incident pulse in pipe 3 has a pressure ratio, Pi3, of 0.8 and 1.1, respectively.
The output data for the calculations using the "constant pressure" theory are shown in
Table 2.14.2. The usual symbols, Pr,, Pr2, and Pr3' are used for the pressure ratios ofthe three
reflected pressure waves at the branch; and c1, c2, and C3 are used for the superposition veloci-
ties. The output data obtained when the more complex theory is employed are shown in Table
2.14.3. The computed mass flow rates (in g/s units) fill, r2, and mi3, are shown in Table 2.14.4.
Table 2.14.5 shows the "errors" in the computed mass flow between the "constant pressure"
234
TABLE 2.14.1 INPUT DATA FOR THE CALCULATIONS

No. Pi1 P12 P13 dl d2 d3 012 013 toi
1 1.4 1.0 1.0 25 25 25 30 180 20
2 1.4 1.0 1.0 25 25 35 30 180 20
3 1.4 1.0 0.8 25 25 25 30 180 20
4 1.4 1.0 1.1 25 25 25 30 180 20
TABLE 2.14.2 OUTPUT DATA FROM THE CALCULATIONS USING

"CONSTANT PRESSURE" THEORY
No. Pr, Pn Pr C1 C2 C3
1 0.891 1.254 1.254 112.6 -56.3 -56.3
2 0.841 1.188 1.188 126.3 -42.7 -42.7
3 0.767 1.086 1.345 148.5 -20.4 -128.1
4 0.950 1.333 1.215 97.0 -72.0 -25.0
TABLE 2.14.3 OUTPUT DATA FROM THE CALCULATIONS USING COMPLEX THEORY
No. Pn. Pr2 P cl C2 C3
1 0.9073 1.2178 1.2760 108.2 -49.0 -60.8
2 0.8481 1.1646 1.1967 124.4 -37.7 -44.6
3 0.7680 1.0819 1.3482 148.0 -19.4 -128.6
4 0.9796 1.2734 1.2517 89.5 -60.3 -32.4
TABLE 2.14.4 OUTPUT ON MASS FLOW FROM CONSTANT PRESSURE

AND COMPLEX THEORY
CONSTANT PRESSURE THEORY COMPLEX THEORY
No. mXl m Ml 2m
1 78.3 -39.1 -39.1 76.0 -33.3 -42.8
2 84.5 -28.5 -55.9 83.6 -24.9 -58.7
3 93.2 -12.8 -80.4 93.0 -12.1 -80.9
4 70.4 -52.3 -18.1 66.4 -42.4 -24.0
TABLE 2.14.5 FURTHER OUTPUT DATA REGARDING ERRORS ON MASS FLOW

No. m, error % M2 error % M3 error % ITERATIONS
1 3.0 17.4 -8.6 2
2 1.1 14.4 -4.8 2
3 0.2 5.7 -0.6 2
4 6.0 23.3 -24.6 2
235
theory and the more complex theory. The final column of Table 2.14.5 shows the number of
iterations needed for the solution; the fact that only two iterations were required for a worst-
case error of 0.05% on any variable reveals the true value of the Benson "constant pressure"
theory in providing a very high-quality first guess for the "unknowns."
It can be seen in test 1 that the constant pressure theory takes no account of the branch
angle, nor of the non-isentropic nature of the flow, and this induces mass flow differences of
up to 17.4% compared to the more complex theory. The actual values of the reflected pressure
waves are quite close for both theories, but the ensuing mass flow error is significant-and is
a telling argument for the inclusion of the more complex theory in any engine simulation
method requiring accuracy.
In test 2 the results for the two theories are closer, i.e., the mass flow errors are smaller, an
effect that occurs by virtue of the fact that pipe 3 has a larger diameter of 35 mm, which
reduces the particle velocity into pipe 2. The pressure loss around the intersection into pipe 2,
which is angled back from pipe 1 at 30°, is seen from Eq. 2.14.1 to be a function ofthe square
of the superposition velocity, c2 . This decreases the pressure loss error within the computa-
tion and in reality. s2
This effect is exaggerated in test 3 where, even though the pipe diameters are equal, the
suction wave incident at pipe 3 likewise reduces the gas particle velocity entering pipe 2. Here,
the errors in mass flow are reduced to a maximum of only 5.7%.
The reverse effect is shown in test 4 where an opposing compression wave incident at the
branch in pipe 3 forces more gas into pipe 2. The mass flow errors now rise to a maximum
value of 24.6%.
In these tests comparing simple and complex theory, the amplitudes of the reflected pres-
sure waves are quite similar. However, the compounding effect of any error on pressure settles
on the density, and the non-isentropic nature of the flow alters the particle velocities, all of
which gives rise to the more serious errors in the mass flow rate as computed by the "constant
pressure" theory.
2.15 Reflection of Pressure Waves in Tapered Pipes
The presence of tapered pipes in the ducts of an engine is commonplace. The action ofthe
tapered pipe in providing pressure wave reflections is often used as a tuning element to signifi-
cantly enhance the performance of the engine. The fundamental reason for this effect is that
the tapered pipe acts as either a nozzle or as a diffuser, in other words acts to produce a more
gradual process of reflection ofpressure waves experienced abruptly at the sudden expansions
and contractions discussed in Secs. 2.10 and 2.11. Almost by definition the process is not only
more gradual but is more efficient at reflecting wave energy because it is spread out in terms of
time. As a consequence, any ensuing tuning effect on the engine is not only more pronounced
but is effective over a wider speed range.
Because a tapered pipe acts to produce a gradual and continual process of reflection, irre-
spective of whether the pipe area is increasing or decreasing, it must be analyzed in a similar
fashion. The ideal would be to conduct the analysis in very small distance steps over the
tapered length, but this would be impractical because it would be a very time-consuming
process-as indeed CFD is!
236
A practical method of analyzing the geometry of tapered pipes is shown in Fig. 2.15. For
simplicity of treatment, I will assume that the pipes are circular in section, but the translation
ofthe analytic principle to ducts ofany cross-section profile will become obvious. The length
L for the section or sections to be analyzed is usually selected to be compatible with the rest of
the computation process for the ducts ofthe engine [2.31]. The tapered section ofthe pipe has
a taper angle 0, which is the included angle of that taper. Having selected a length L over
which the unsteady gas dynamic analysis is to be conducted, it is a matter of simple geometry
to determine the diameters at the various locations on the tapered pipe. Consider sections 1
and 2 in Fig. 2.15. They are of equal length L. At the commencement ofsection 1, the diameter
is da, and at its conclusion it is db. At the start ofsection 2, the diameter is db, and it is dc at its
conclusion.
Any reflection process for sections 1 and 2 is considered to take place at the interface as a
"6sudden" expansion or contraction, depending on whether the particle flow is acting in a dif-
fusing manner as in Fig. 2.15(b) or in a nozzle fashion as in Fig. 2.15(c). In short, the flow
proceeds in an unsteady gas dynamic process along section 1 in a parallel pipe of representa-
tive diameter, dl, with area, Al, and is then reflected at the interface to section 2, where the
representative diameter is d2, with area, A2. This is the analytic case irrespective of whether
the flow is acting in a diffusing manner as in Fig. 2.15(b) or in a nozzle fashion as in Fig.
2.15(c). The logical mean diameter, or mean area, for each of the sections is the mean area
corresponding to the tapered volume within each section. Assuming a round pipe, the section
volume, V, is the frustum of a cone. Ifthe tapered section is irregular, or non-circular, then that
geometry will have to be determined, but the principle is the same. The section volumes, V1
and V2, for tapered round pipes are:
V 12-d+dadb+db and V2 12 d +dbdc+d) (2.15.1)
The mean diameters for each conical section are related to the mean areas and actual
section volumes by:
Al = V hence d= 4A and A2= L2hence d2 = (2.15.2)
The analysis of the flow commences by determining the direction of the particle flow at
the interface between section 1 and section 2 and the area change occurring at that position. If
the flow is behaving as in a diffuser, then the ensuing unsteady gas dynamic analysis is con-
ducted using the theory precisely as presented in Sec. 2.10 for sudden expansions. Ifthe flow
is behaving as in a nozzle, then the ensuing unsteady gas dynamic analysis is conducted using
the theory precisely as presented in Sec. 2.11 for sudden contractions.
237
dc
da
I
db
I I
|
is
X\ ,
s
s
QO
I
Z
J
I
-
I I
1< L L
(a) the dimensioning of the tapered pipe
(b) flow in a diffuser (c) flow in a nozzle

Fig. 2.15 Treatment of taperedpipesfor unsteady gas-dynamic analysis.
2.15.1 Separation of the Flowfrom the Walls of a Di/fiuser

One of the issues that is always debated in the literature is flow separation from the walls
of a diffuser, the physical situation being as in Fig. 2.15(b). Under such circumstances, the
flow detaches from the walls into a central, highly turbulent core. As a consequence, the
entropy gain is proportionately higher than in the thermodynamics shown in Fig. 2.9(a), for
the pressure drop is not so great and the temperature drop is also lessened due to energy dissi-
pation in turbulence. It is postulated that in such a flow separation the process becomes almost
isobaric and can be represented as such in the thermodynamic analysis ofSec. 2.10. Therefore,
if flow separation in a diffuser is assumed to occur, the analytic process set out in Sec. 2.10
should be amended to replace the equation that tracks the non-isentropic flow in the normal
attached mode, namely the momentum equation, with another equation that better simulates
the greater entropy gain of separated flow, namely a constant pressure equation.
Hence, in Sec. 2.10, Eq. 2.10.4 (or Eq. 2.10.9 in its final format) should be deleted from
the set of equations to be analyzed, and it should be replaced with Eq. 2.15.3 (or its more
detailed equivalent, Eq. 2.15.4) below.
238
The assumption is that the particle flow is moving, and diffusing, from section 1 to section
2 as in Fig. 2.15(b), and that wall separation has been detected. Constant superposition pres-
sure at the interface between section 1 and 2 produces the following function, using the same
variable nomenclature as in Sec. 2.10.
I Ps2 = 0
PsF (2.15.3)
This "constant pressure" equation is used to replace the final form of the momentum equa-
tion in Eq. 2.10.9. The "constant pressure" equation can be restated in the form below and used
in the computational process:
x G7-x_ 7 =0 (2.15.4)
You may well inquire at what point in a computation should this change of analytical tack
be implemented? In many gas dynamics texts, where steadyflow is being described, either
theoretically or experimentally, the conclusion reached is that flow separation will take place
if the particle Mach number is greater than 0.2 or 0.3 and, more significantly, if the included
angle, 0, of the tapered pipe is greater than a critical value, reported widely in the literature as
lying between 5 and 7°. The work to date at QUB indicates that the angle, 0, is of very little
significance, but that gas particle Mach number alone is the important factor controlling flow
separation. Phrased mathematically, the current conclusion would be:
if MsiI 0.65 then use the constant pressure equation, Eq. 2.15.4. (2.15.5)
if Ms, < 0.65 then use the momentum equation, Eq. 2.10.9 (2.15.6)
Future work in correlating theory with experiment, such as that seen in Sec. 2.19.7, may
shed more light on this subject. Suffice it to say that there is enough evidence already to
confirm that any computational method that universally employs the momentum equation for
the solution of diffusing flow in steeply tapered pipes where the Mach number is high will
inevitably produce a very inaccurate assessment of the unsteady gas flow behavior.
2.16 Reflection of Pressure Waves in Pipes for Outflow from a Cylinder
This situation is fundamental to all unsteady gas flow generated in the intake or exhaust
ducts of a reciprocating ic engine. Fig. 2.16 shows an exhaust valve (or port) and pipe, or the
throttled end of an exhaust pipe leading into a plenum such as the atmosphere or a silencer
box. Anywhere in an unsteady flow regime where a pressure wave in a pipe is incident on a
pressure-filled space, box, plenum, or cylinder, the following method is applicable to deter-
mine the magnitude of the mass outflow, of its thermodynamic state, and of the reflected
239
AsumtinsPt P2
clAO Tt T2
A1>>A2 Pt P2
£.1 N At eff A2
-30.
P1 41-
Cl -.4 . Ti2
AIM
yl
11-i 30.
Ri S Pr2 -
-0.
Fig. 2.16 Outflowfrom a cylinder or plenum to a pipe.
pressure wave. The theory to be generated is generally applicable to an intake valve (or port)
and pipe for inflow into a cylinder, plenum, crankcase, or at the throttled end of an intake pipe
from the atmosphere or a silencer box. However, the subtle differences for this analysis are
given separately in Sec. 2.17.
You may well be tempted now to ask what then is the difference between this theoretical
treatment and that given for the restricted pipe scenario in Sec. 2.12, for the drawings in Figs.
2.16 and 2.12 look remarkably similar. The answer is direct: In the theory presented here, the
space from which the particles emanate is considered to be sufficiently large and the flow so
three-dimensional as to give rise to the fimdamental assumption that the particle velocity within
the cylinder is considered to be zero, i.e., cl is zero.
The solution of the gas dynamics of the flow must include separate treatments for sub-
sonic outflow and sonic outflow. The first presentation ofthe solution for this type of flow was
by Wallace and Nassif [2.5], and their basic theory was used in a computer-oriented presenta-
tion by Blair and Cahoon [2.6]. Probably the earliest and most detailed exposition ofthe deri-
vation of the equations involved is that by McConnell [2.7]. However, whereas all of these
presentations declared that the flow was analyzed non-isentropically, a subtle error was intro-
duced within the analysis which negated that assumption. Moreover, all of the earlier
solutions, including that by Bingham [2.19], used fixed values ofthe cylinder properties through-
out and solved the equations with either the properties of air (y = 1.4 and R = 287 J/kgK) or
exhaust gas (y = 1.35 and R = 300 J/kgK). The arithmetic solution was stored in tabular form
and indexed during the course of a computation. Today, that solution approach is inadequate,
for the precise equations in fully non-isentropic form must be solved at each instant of a com-
putation for the properties of the gas existing at that location, at that juncture.
Because a more complex solution, i.e., that for restricted pipes in Sec. 2.12, has already
been presented, the complete solution for outflow from a cylinder or plenum in an unsteady
gas dynamic regime will not pose any new theoretical difficulties.
240
The case of subsonic particle flow will be presented first and that for sonic flow is pre-
sented in Sec. 2.16.1.
Subsonic Ou4low
In Fig. 2.16 the expanding flow from the throat to the downstream superposition point 2 is
seen to leave turbulent vortices in the corners of that section. That the streamlines ofthe flow
give rise to particle flow separation implies a gain of entropy from the throat to area section 2.
On the other hand the flow from the cylinder to the throat is contracting and can be considered
to be isentropic in the same fashion as the contractions discussed in Secs. 2.11 and 2.12. This
is summarized on the temperature-entropy diagram in Fig. 2.17, where the gain of entropy for
the flow rising from pressure pt to pressure Ps2 is clearly visible. The isentropic nature of the
flow from Pi to pt-it is a vertical line on Fig. 2.17-can also be observed.
The properties and composition of the gas particles are those of the gas at the exit of the
cylinder to the pipe. The word "exit" is used most precisely. For most cylinders and plenums,
the process of flow within the cylinder is one of mixing. In such a case the properties of the gas
at the exit from it for an outflow process is that of the mean of all of the contents. Not all
internal cylinder flow is like that. Some engine cylinders have a stratified in-cylinder flow
process. A two-stroke engine cylinder during scavenging would be a classic example of this
situation. There, the properties of the gas exiting the cylinder would vary from combustion
products only at the commencement of the exhaust outflow to a gas containing increasingly
larger proportions of the air lost during the scavenge process; it would be mere coincidence if
the exiting gas at any instant had the same properties as the average of all of the cylinder
contents.
It is not inconceivable that the exhaust outflow process during the valve overlap period of
the four-stroke engine is similar. This is illustrated in Fig. 2.25 where there are stratified zones
labeled as CX surrounding the intake and exhaust apertures. The properties ofthe gas in those
zones may well differ from the mean values for all of the cylinder, labeled in Fig. 2.25 as Pcl
Tc, etc. and also the gas properties RC and yc. In such a case a means of tracking the extent of
the stratification must be employed and these variables determined as PCXg TcxS RCXS ycx, etc.,
and employed for those properties that are subscripted with a 1 in the text below. Further
debate on this issue is found in Sec. 2.18.10.
Although this singularity of stratification should always be borne in mind, and dealt with
should it arise, the various gas properties for cylinder outflow are defined as:
y-=yj R= R1 G5 =G51 G7=G71,etc.

continuity, the First Law of Thermodynamics, and the momentum equation.
Density Poi = Pot = Po P02 -P (2.16.1)

RTO RT02
241
W ISENTROPIC LINE P1
a:
!R;
I- / PS2
w P
a.T i
- @ ENTROPY
(a) temperature-entropy characteristics for subsonic outflow.
W
a: ISENTROPIC LINE P1
H
a:
w s
w
Ht
(b) temperature-entropy characteristics for sonic outflow.

Fig. 2.17 Temperature-entropy characteristics for cylinder orplenum outflow.
Acoustic velocity ao, = aot = yRTo0 aO2= yRT02 (2.16.2)
The continuity equation for mass flow in previous sections is still generally applicable and
repeated here, although the entropy gain is reflected in the reference acoustic velocity and
lilt - m2 0 (2.16.3)
242
The convention for particle flow direction is not significant, and this equation becomes:
pt(CcAt)(Csct) - P02XG5A2G5ao2(Xi2 - Xr2) = 0 (2.16.4)

The above equation, for the mass flow continuity for flow from the throat to the down-
stream station 2, contains the coefficient of contraction on the flow area, Cc, and the coeffi-
cient of velocity, Cs. These are conventionally connected in fluid mechanics theory to a
coefficient of discharge, Cd, to give an effective throat area, Ateff, as follows:
Cd = CcCs and Ateff = CdAt

Chapter 3 is devoted to a full discussion on discharge coefficients, not only for poppet
valves but for all types of ports and pipe end restrictions.
The equation of mass flow continuity, Eq. 2.16.4, becomes:
CdptAtct- P2XAs2 2G5aO2 (Xi2 Xr2) = 0 -
for flow from the cylinder to superposition station 2 can be expressed as:
2l 2= hs2 + C22s2
2
h, +
or G5a2 -(G as2 +C22)0 (2.16.5)

The First Law of Thermodynamics for flow from the cylinder to the throat can be
expressed as:
2 2
+Cl = ht + ct
h
o+2 2
or CP(TI -To) 2L= 0 (2.16.6)
243
The momentum equation for flow from the throat to superposition station 2 is expressed
as:
A 2(Pt -Ps2N) + s2(Ct - Cs2) =O (2.16.7)

The unknown values will be the reflected pressure wave at the boundary, Pr2, the reference
temperature at position 2, T02, the throat pressure, Pt, and the velocity ct, at the throat. This
gives four unknowns, necessitating four equations, namely the mass flow equation in Eq.
2.16.4; the two First Law equations, Eq. 2.16.5 and Eq. 2.16.6; and the momentum equation,
Eq. 2.16.7. All other "unknown" quantities can be computed from these values and from the
"known" values. The known values are the downstream pipe area, A2, the throat area, At, the
gas properties exiting the cylinder, and the incident pressure wave, Pi2.
Recalling that,
C G17 G17 ( G17

X J=Pi
P
and Xi2= Po
and setting Xt
t
Pt
Po
then due to isentropic flow from the cylinder to the throat, the temperature reference condi-
tions are given by:
a, =aOXI or To- 2
Because T1 and X1 are input parameters to any given problem, To, is readily determined.
In such a case, from Eqs. 2.16.1 and 2.16.2, the reference densities and acoustic velocities for
the cylinder and throat conditions are also readily determined. As shown below, the density
and temperature at the throat can also be related to the reference conditions.
=
p01XG5 and Tt =
Pt -Pt PoixtyR
The continuity equation set in Eq. 2.16.4 reduces to:
PolXt CdAtct-Po2(Xi2 +Xr2-l)GA2G5aO2(Xi2-Xr2)=O (2.16.8)
244
G5(aOXj)2 5[(GaO2(Xi2- Xr2)) +G5ao2(Xi2 +Xr2 1)2] (2.16.9)
G5[(aoIXl)2 -(aoixt )2J-c =0 (2.16.10)
The momentum equation Eq. 2.16.7 reduces to:
Po[xtG7 _(Xi2 + Xr2 1)G7

+ P02(Xi2 + Xr2 1)G5 xG5a02(Xi2 Xr2)] -
(2.16.11)
x[ct - G5aO2(Xi2 - Xr2)I = 0
The four equations, Eqs. 2.16.8 to 2.16.11, cannot be reduced any further because they are
polynomial functions ofthe four unknown variables, Xr2,, Xt, a02, and ct. These functions can
be solved by a standard iterative method for such problems. I use the Newton-Raphson method
for the solution of multiple polynomial equations because it is stable, accurate, and rapid in
execution. The arithmetic solution on a computer is conducted by a Gaussian Elimination
method.
2.161 Outflowfrom a Cylinder Where Sonic Pardcle Velocity is Encountered
In the above analysis of unsteady gas outflow from a cylinder the particle velocity at the
throat will quite commonly be found to reach, or even attempt to exceed, the local acoustic
velocity. In thermodynamic or gas dynamic terms, this is not possible. The highest particle
velocity that is permissible is the local acoustic velocity at the throat, i.e., the flow is permitted
to become choked. Therefore, during the mathematical solution of Eqs. 2.16.8 to 2.16.11, the
local Mach number at the throat is monitored and retained at unity if it is found to exceed it.
As, Mt = t =-1 then ct = aOIXt (2.16.12)

aoiXt
245
With ct now fixed as a function of Xt, the First Law of Thennodynamics for outflow from
the cylinder to the throat, in Eq. 2.16.10, now gives a direct solution for the pressure ratio from
the cylinder to the throat. The combination of Eqs. 2.16. 10 and 2.16.12 provides:
G5{(aolXl) - (aoiXt ) - (aoiXt )2 = 0
) 3
(2
=_Pt
___ _
Consequently, Xo o
P [Y+) (2.16.13)
The pressure ratio from the cylinder to the throat where the flow at the throat is choked,
i.e., where the Mach number at the throat is unity, is known as "the critical pressure ratio." The
derivation of this ratio is also to be found in many standard texts on thermodynamics or gas
dynamics. It is only applicable ifthe upstream particle velocity is considered to be zero. Con-
sequently it is not a universal "law" and it must be applied only where the thermodynamic
assumptions used in its creation are relevant. Significantly, it is not to be found in either Secs.
2.12.1 or 2.17.1.
This simplifies the sonic outflow case because it gives a direct solution for two of the
unknowns and replaces two of the four equations in the subsonic solution. From an arithmetic
standpoint, it is probably simpler to eliminate the momentum equation, and use only the con-
tinuity equation and the First Law of Eqs. 2.16.8 and 2.16.9, but from a thermodynamic stand-
point, it is more accurate to retain it!
The acquisition of all related data for pressure, density, particle velocity and mass flow
rate at both superposition stations and at the throat follows directly from the solution of the
remaining two polynomial equations for Xr2 and aO2.
2.16.2 Numerical Examples of Outflowfrom a Cylinder
The application of the above theory is illustrated by the calculation of outflow from a
cylinder using the data given in Table 2.16.1. The nomenclature for these data is consistent
with the theory and the associated sketch in Fig. 2.17. The units of the data, if inconsistent with
strict SI units, are indicated in the several tables. The calculation output is shown in Tables
2.16.2 and 2.16.3.
The input data for tests numbered 1 and 2 pertain to a "blowdown" situation from gas at
high temperature and pressure, with a small diameter port simulating a cylinder valve (or port)
that has just commenced opening. The cylinder has a pressure ratio of5.0 and a temperature of
1000°C. The exhaust pipe diameter, at 30 mm, is the same for all of the tests. In tests 1 and 2
the valve area is equivalent to a 3-mm diameter hole and has an assumed coefficient of dis-
charge of 0.90. The gas in the cylinder and in the exhaust pipe in test 1 has a purity of zero, i.e.,
it is all exhaust gas.
246
TABLE 2.16.1 INPUT DATA TO CALCULATIONS OF OUTFLOW FROM A CYLINDER

No. P1 T1 °C -II dt mm d2 mm Cd P12 rI2
1 5.0 1000 0.0 3 30 0.9 1.0 0.0
2 5.0 1000 1.0 3 30 0.9 1.0 1.0
3 1.8 500 0.0 25 30 0.75 1.0 0.0
4 1.8 500 0.0 25 30 0.75 1.1 0.0
5 1.8 500 0.0 25 30 0.75 0.9 0.0
TABLE 2.16.2 OUTPUT FROM CALCULATIONS OF OUTFLOW FROM A CYLINDER

No. Pn Pq2 Ts2 °C Pt Tt OC m, 9/s
1 1.0351 1.0351 999.9 2.676 805.8 3.54
2 1.036 1.036 999.9 2.641 787.8 3.66
3 1.554 1.554 486.4 1.319 440.0 85.7
4 1.528 1.672 492.5 1.546 469.5 68.1
5 1.538 1.392 479.9 1.025 392.9 94.3
TABLE 2.16.3 FURTHER OUTPUT FROM CALCULATIONS OF

OUTFLOW FROM A CYLINDER
No. Ct Mt Cs2 M2 abm and aot ao2
1 663.4 1.0 18.25 0.025 582.4 717.4
2 652.9 1.0 18.01 0.025 568.3 711'.5
3 372.0 0.69 175.4 0.315 519.6 525.1
4 262.9 0.48 130.5 0.234 519.6 522.1
5 492.7 0.945 213.5 0.385 519.6 530.5
The purity, rI, defines the gas as being a mixture of air and exhaust gas, where the air is
assumed to have a specific heats ratio, y, of 1.4 and a gas constant, R, of 287 J/kgK. The
exhaust gas is assumed to have a specific heats ratio, y, of 1.36 and a gas constant, R, of 300
J/kgK. For further explanation see Eqs. 2.18.47 to 2.18.50.
In test 1, where the cylinder gas is assumed to be exhaust gas, the results of the computa-
tion are presented in Tables 2.16.2 and 2.16.3, and show that the flow at the throat is choked,
i.e., that Mt is 1.0. It is shown in the table that a small pulse with a pressure ratio ofjust 1.0351
is sent into the exhaust pipe. The very considerable entropy gain is evident by the disparity
between the reference acoustic velocities at the throat and at the pipe, a%t and a02, at 582.4 and
717.4 m/s, respectively. It is clear that any attempt to solve this flow regime as an isentropic
process would be very inaccurate.
247
Design and Simulation ofFouk-Stroke Engines
The presentation here of a non-isentropic analysis with variable gas properties is unique
and its importance can be observed by comparing the results of tests 1 and 2. The cylinder gas
in test data set 1 is exhaust gas. Test data set 2 is identical to set 1, except that the purity in the
cylinder and in the pipe is assumed be unity, i.e., it is air. The mass flow rate calculated for data
set 1 is 3.54 g/s and it is 3.66 g/s for data set 2; that is, there is an error of 3.4%. Mass flow
errors in simulation translate ultimately into errors in the prediction of air mass trapped in a
cylinder, the value of which is directly related to power output. This error of 3.4% is actually
more significant than it appears because the effect is compounded throughout the entire engine
cycle during the computer simulation.
Test data sets 3 to 5 illustrate the ability of pressure wave reflections to influence dramati-
cally the "breathing" of an engine. The situation is one of exhaust flow from a cylinder with
gas at high temperature and pressure. A large-area port simulates a valve that is at a well-open
position. The cylinder gas has a pressure ratio of 1.8 and a temperature of 500°C. The exhaust
pipe diameter, at 30 mm, is the same for all ofthe tests. The valve area is equivalent to a 25 mm
diameter hole and has a typical coefficient of discharge of 0.75. The gas in the cylinder and in
the exhaust pipe has a purity ofzero, i.e., it is all exhaust gas. The only difference between data
sets 3 to 5 is the amplitude of the pressure wave in the pipe incident on the exhaust port. In data
set 3 it has a pressure ratio of 1.0, or undisturbed conditions exist; in data set 4 it has a pressure
ratio of 1.1, providing a modest opposition to the outflow; in data set 5 it has a pressure ratio of
0.9, giving a modest suction effect on the cylinder. The results show very considerable
variations in the ensuing mass flow rate exiting the cylinder, ranging from 85.7 g/s when the
conditions are undisturbed in test 3, to 68.1 g/s when the incident pressure wave is of one
compression, to 94.3 g/s when the incident pressure wave is one of expansion. These swings of
mass flow rate represent variations of -20.5% to +10%. It will be observed that test 4 with the
lowest mass flow rate, has the highest superposition pressure ratio, PS2, at the pipe point, and
that test 5 with the highest mass flow rate, has the lowest superposition pressure in the pipe.
Because this pressure, Ps2, is that measured by a fast-response pressure transducer, you might
be tempted now to conclude that test 3 is the one with the stronger wave action. Such is the
folly of casually examining measured pressure traces in the exhaust ducts of engines-but this
is nothing new to those who have studied Sec. 2.2.1!
This illustrates perfectly both the advantages of using pressure wave effects in the exhaust
system of an engine to enhance the mass flow through it and the disadvantages of poorly
designing the exhaust system. These simple numeric examples reinforce the opinions expressed
earlier in Sec. 2.8.1 regarding the effective use of reflections of pressure waves in exhaust
pipes. There is more ofthat ilk in Chapters 5 and 6.
2.17 Reflection of Pressure Waves in Pipes for Inflow to a Cylinder
This situation is fundamental to all unsteady gas flow generated in the intake or exhaust
ducts of a reciprocating ic engine. Fig. 2.18 shows an inlet valve (or port) and pipe, or the
throttled end of an intake pipe leading into a plenum such as the atmosphere or a silencer box.
In an unsteady flow regime anywhere a pressure wave in a pipe is incident on a pressure filled
space, box, plenum or cylinder, the following method is applicable to determine the magnitude
of the mass inflow, of its thermodynamic state, and of the reflected pressure wave.
248
Ps2
CYLINDER Pt s2
P1 A
Pi
Al
Pr2 -
Fig. 2.18 Inflowfrom a pipe to a cylinder or a plenum.
In the theory presented here, the space into which the particles disperse is considered to be
sufficiently large and also three-dimensional as to permit the fundamental assumption that the
particle velocity within the cylinder is considered to be zero.
cl =0 (2.17.1)
The case of subsonic particle flow will be presented first and that for sonic flow is pre-
sented in Sec. 2.17.1.
Subsonic Inflow
In Fig. 2.18, the expanding flow from the throat to the cylinder produces pronounced
turbulence within the cylinder. The taditional assumption is that this dissipation of turbulence
energy produces no pressure recovery from the throat ofthe port or valve to the cylinder. This
assumption applies only where subsonic flow is maintained at the throat.
Pt =Pi (2.17.2)
On the other hand, the flow from the pipe to the throat is contracting and can be considered
to be isentropic in the same fashion as the other contractions discussed in Secs. 2.11 and 2.12.
This is summarized on the temperature-entropy diagram in Fig. 2.19, where the gain of
entropy for the flow rising from pressure Pt to cylinder pressure Pj is clearly visible. The
isentropic nature of the flow from Ps2 to pt-it is a vertical line on Fig. 2.19-can also be
observed.
249
W ISENTROPIC LINE Ps2

H
D*
cL T2 2:1 1
w-------------------- P
IT1
0-1 M----
---- Tt
02
ENTROPY
(a) temperature-entropy characteristics for subsonic inflow.
W ISENTROPIC LINE s2
w P
- @ ENTROPY
(b) temperature-entropy characteristics for sonic inflow.
Fig. 2.19 Temperature-entropy characteristics for cylinder or plenum inflow.
The properties and composition of the gas particles are those of the gas at the superposi-
tion point in the pipe. The various gas properties for cylinder inflow are defined as:
Y=Y2 R=R2 G5=G52 G7=G72,etc.

continuity, the First Law of Thermodynamics, and the momentum equation. However, the
momentum equation is not used in this particular analysis for subsonic inflow, because the
constant pressure assumption used in Eq. 2.17.2 reflects an even higher gain of entropy, i.e.,
energy dissipation due to turbulence, than would be the case if the momentum equation were
250
to be involved. The constant pressure assumption, in connection with a large entropy gain, has
been used in this regard before within this text, notably in the section dealing with separation
of diffusing flow in tapered pipes (Sec. 2.15.1).
Density P02=Pot = Poi -T (2.17.3)

RT02 RTo,
Acoustic velocity aO2 = aot =yRT ao0 =yRT0 (2.17.4)
The continuity equation for mass flow between the pipe and the throat is:
nt -rn2=0 (2.17.5)
This equation becomes:
Pt(CcAt)(Csct) - P02X GA2G5a02(Xi2 - Xr2) = 0 (2.17.6)
The above equation, for the mass flow continuity for flow from the upstream station 2 to
the throat t, contains the coefficient of contraction on the flow area, Cc, and the coefficient of
velocity, C.. These are conventionally connected in fluid mechanics theory to a coefficient of
discharge, Cd, to give an effective throat area, Aftff, as follows:
Cd = CcCs and Ateff = CdAt

Discharge coefficients are discussed at length in Chapter 3. The mass flow continuity
equation becomes:
CdptAtco - p02XsA2G5ao2(Xi2 - Xr2) = 0

The First Law of Thermodynamics for flow from the pipe to the throat can be expressed
as:
h2 + 2 =ht + 2
o2 2
or Cp(T2 -
Tt( c22 C-__2 (.77
251
The unknown values will be the reflected pressure wave at the boundary, Pr2, and the
velocity, ct, at the throat. There are two unknowns, necessitating two equations, namely the
mass flow equation in Eq. 2.17.6, and the First Law equation, Eq. 2.17.6. All other "unknown"
quantities can be computed from these values and from the "known" values. The known val-
ues are the downstream pipe area, A2, the throat area, At, the gas properties at superposition
station 2 in the pipe, and the incident pressure wave, Pi2. You will recall that:
PG17 G17 G17 PG17

X = J and X 2 =Pi2 andsetting Xt = Pt =p
Po) Po J Po PoJ
The reference temperature ToI for the cylinder is given by:
a, = aOXI or To0 =-
2
Because T1 and X1 are input parameters to any given problem, ToI is readily determined.
In such a case, from Eqs. 2.17.1 and 2.17.2, the reference densities and acoustic velocities for
the cylinder and throat conditions are also readily determined. The pipe point and the throat
point are linked isentropically. Consequently, the density and temperature at the throat are
related to their isentropic reference conditions and to the assumption regarding the throat pres-
sure equality with cylinder pressure.
G5 G5 (ao2Xi)
Pt = P2Xt =P02X1 an T yR
Hence, the continuity equation set in Eq. 2.17.6 reduces to:
c P-p02(Xi2 + Xr2 - 1) A2G5aO2(Xi2 -Xr2) =0

pO2Xl-CdA (2.17.8)
and the First Law of Thermodynamics in Eq. 2.17.7 reduces to:
5(aO2(Xi2 + Xr2 - 1))2 -G5(ao2X1)2]

+[(G5aO2(Xi2 - - ct] =0 (2.17.9)
252
Eqs. 2.17.8 and 2.17.9 cannot be reduced any further as they are polynomial functions of
the two variables, Xri and ct. These functions can be solved by a standard iterative method for
such problems. I have determined that the Newton-Raphson method for the solution of mul-
tiple polynomial equations is stable, accurate, and rapid in execution. The arithmetic solution
on a computer is conducted by a Gaussian Elimination method.
2.17.1 Inflow to a Cylinder Where Sonic Paricle Velocy is Encountered
In the above analysis of unsteady gas outflow from a cylinder, the particle velocity at the
throat will quite commonly be found to reach, or even attempt to exceed, the local acoustic
velocity. In thermodynamic or gas dynamic terms, this is not possible. The highest particle
velocity that is permissible is the local acoustic velocity at the throat, i.e., the flow is pennitted
to become choked. Therefore, during the mathematical solution of Eqs. 2.17.8 and 2.17.9, the
local Mach number at the throat is monitored and retained at unity if it is found to exceed it.
As, Mt = =I then ct = aO2Xt

-1 (2.17.10)
aO2Xt
The solution for a critical pressure ratio, as in Sec. 2.16.1, cannot be used here because the
upstream particle velocity in the pipe is not zero.
Eq. 2.17.10 simplifies the entire procedure for this and gives a direct relationship between
two of the unknowns. However, this does not reduce the solution complexity because it also
eliminates a previous assumption for the subsonic flow regime that the throat pressure is equal
to the cylinder pressure, i.e., the assumption made in Eq. 2.17.2. Therefore, Eqs. 2.17.8 and
2.17.9, must be revisited and that assumption removed.
After this revision, the continuity equation in Eq. 2.17.8 becomes:
aO2PO2XG6CdAt P02(Xi2 + Xr2 - 1)G5 A2G2a12(X12 - Xr2) =0° (2.1)

-
After this revision, the First Law of Thermodynamics in Eq. 2.17.9 becomes:
[G5(aO2(Xi2 + Xr2 -1))2 _G5(ao2Xt)2]

+ 5a02(Xi2 - Xr2 - (aO2Xt )2] = 0 (2.17.12)
Eqs. 2.17.11 and 2.17.12, cannot be reduced any further because they are polynomial
functions of the same two variables, Xr1 and Xt. These functions can be solved by the same
standard iterative method as used in the case of subsonic inflow.
253
rate at the pipe superposition station and at the throat follows directly from the solution ofthe
two polynomials for Xr2 and Xt.
2.17.2 Numerical Examples ofInflow into a Cylinder
The application of the above theory is illustrated by the calculation of inflow into a cylin-
der using the data given in Table 2.17. 1. The nomenclature for the data is consistent with the
theory and the associated sketch in Fig. 2.18. The units ofthe data, if inconsistent with strict SI
units, are indicated in the several tables. The calculation output is shown in Tables 2.17.2 and
2.17.3.
The input data for all of the tests are with reference to a nonnal intake situation from
atmospheric air through a well-open valve, equivalent in area to a 20 mm diameter hole with a
discharge coefficient of0.75, simulating a cylinder valve or valve that is near to its maximum
opening. The cylinder has a pressure ratio of 0.65 and a temperature of 70°C. The reference
temperature at the pipe point is 70°C, an apparently high temperature that would not be abnor-
mal for an intake system, the walls ofwhich have been heated by conduction from the rest of
the power unit. The intake pipe diameter, at 30 mm, is the same for all of the tests. The gas in
the cylinder and in the induction pipe has a purity of 1.0, i.e., it is air.
Test data sets 1, 2, and 3 illustrate the ability of pressure wave reflections to influence the
"breathing" of an engine. The difference between data sets 1 to 3 is the amplitude of the
pressure wave in the pipe that is incident on the intake valve; in data set 1 it is at a pressure
ratio of 1.0, or undisturbed conditions; in data set 2 it is at a pressure ratio of0.9, indicating it
has been throttled earlier upstream; in data set 3 it is at a pressure ratio of 1.2, giving a good
ramming effect on the cylinder. The results show very considerable variations in the ensuing
mass flow rate entering the cylinder, ranging from 36.0 g/s when the conditions are undis-
turbed in test 1, to 21.1 g/s when the incident '"trottled" pressure wave is one ofexpansion, to
TABLE 2.17.1 INPUT DATA TO CALCULATIONS OF INFLOW INTO A CYLINDER

No. P1 T02 °C II, dt mm d2 mm Cd P2 II2
1 0.65 70 1.0 20 30 0.75 1.0 1.0
2 0.65 70 1.0 20 30 0.75 0.9 1.0
3 0.65 70 1.0 20 30 0.75 1.2 1.0
4 0.65 40 1.0 20 30 0.75 1.2 1.0
TABLE 2.17.2 OUTPUT FROM CALCULATIONS OF INFLOW INTO A CYLINDER

No. Pn Pv2 Ts2 °C Pt Tt OC Ma2 9/S |
1 0.801 0.801 48.9 0.65 30.3 36.0
2 0.779 0.698 36.5 0.65 30.3 21.1
3 0.910 1.095 78.9 0.65 30.3 57.3
4 0.910 1.095 48.2 0.65 3.75 60.0
254
TABLE 2.17.3 FURTHER OUTPUT FROM CALCULATIONS OF

INFLOW INTO A CYLINDER
No. Ct Mt c2 M42 s,andaot
1 201.9 0.578 58.0 0.161 371.2
2 118.2 0.338 37.4 0.106 371.2
3 321.4 0.921 73.8 0.196 371.2
4 307.0 0.921 70.5 0.196 354.6
57.3 g/s when the incident "ramming" pressure wave is one of compression. These variations
in mass flow rate represent changes of 41.4% to +59.2%. The effect of "intake ramming,"
first discussed in Sec. 2.8.2, has a profound effect on the mass ofair that can be induced by an
engine. In addition, if the intake system is designed badly, and provides expansion wave
reflections at the intake valve during that process, then the potential for the deterioration of
induction of air is also self-evident. (Of course, if the design intention is throttling, then this
criticism is eliminated!)
Test data set 4 is almost identical to test data set 2, except that the intake system has been
cooled, or "intercooled" as in a modern turbo-diesel automobile engine, to give a reference air
temperature at the pipe point of 40°C compared to test data set 2, which held hotter air at 70°C.
The density of the intake air has been increased by 8.8%. This density increase does not trans-
late linearly into the same increase of air mass flow rate. The computation shows that the air
mass flow rate rises from 57.1 to 60.0 g/s, which represents a gain of5.1%. However, because
intake air flow and torque are almost directly related, the charge cooling represented by test
data set 4 indicates a potentially useful gain of torque and power.
2.18 The Simuladon of Engines by the Computation of Unsteady Gas Flow
Many computational methods have been suggested for the solution of this theoretical situ-
ation, such as Riemann variables [2.10], Lax-Wendroff [2.42] and other finite difference
procedures [2.46], and yet others [2.12, 2.49, 2.69]. The basic approach adopted here is to
reexamine the fundamental theory of pressure wave motion and adapt it to a mesh method
intexpolation procedure. At the same time, the boundary conditions for inflow and outflow,
such as the filling and emptying of engine cylinders, are resolved for the generality of gas
properties and in terms of the unsteady gas flow that controls those processes. The same gen-
erality of gas property and composition is traced throughout the pipe system. This change of
gas property is very significant in two-stroke engines where the exhaust blowdown is fol-
lowed by short-circuited scavenge air. For the four-stroke engine, especially high-performance
spark-ignition engines or those with significant valve overlap periods, air can-indeed should-
be drawn into the exhaust system to aid the induction process to attain a high volumetric
efficiency. It is also very significant at varying load levels in diesel engines. Vitally important
in this context is the solution for the continual transmission and reflection of pressure waves
because they encounter both differing temperature gradients and gas properties, and both gradual
and sudden changes of area throughout the engine ducting. Of equal importance is the ability
255
ofthe calculation to predict the effect of intemal heat generation within the duct or of external
heat transfer with respect to it, and to be able to trace the effect of the ensuing gas temperature
change on both the pressure wave system and the net gas flow.
The computation ofunsteady gas flow through the cylinders of reciprocating internal com-
bustion engines is a technology that is now nearly forty years old. The original paper by
Benson et al. [2.10] formalized the use of Rieman variables as a technique for tracing the
motion ofpressure waves in the ducts of engines and the effect on the filling and emptying of
the cylinders attached to them. Subsequently, other models have been introduced using finite
difference techniques [2.29] and Lax-Wendroff methods [2.42, 2.46, 2.49]. Many technical
papers and books have been published on this subject, including some by myself which are
listed in the references.
For a more complete discussion ofthis subject, and in particular for the description ofthe
previous computational procedures produced by authors from the University of Manchester
(UMIST) and QUB, the textbooks by Benson [2.4] and Blair [2.25], and the papers by Chen
et al. [2.47] and Kirkpatrick et al. [2.41] may be studied.
This introduction is somewhat brief, for a full discussion of the computational procedures
developed by others would fill-indeed have filled-many textbooks, much less an introduc-
tion to this chapter. Consequently, many scores of worthy contributors to the literature have
not been cited in the References and I hope that those who are not mentioned here will not feel
slighted in any way.
I started working in the era when the graphical method of characteristics [2.8] was the
only means of computing unsteady gas flow in pipes. While cumbersome in the arithmetic
extreme before the advent ofthe digital computer, and hence virtually ineffective as a design
tool, it had the very considerable merit ofproviding insights into unsteady gas flow in a man-
ner that the Riemann variable, the finite difference, or the Lax-Wendroffmethods, have never
been able to. Doubtless, the programmers of such computational techniques have gained these
insights, but this rarely, if ever, applied to the users of the software, i.e., the engineers and
engine designers who had to employ it. As one who has taught such computational techniques
to several generations of students, I can confirm that the insights gained by students always
occurred most readily from lectures based on the papers produced by Wallace and Nassif [2.5]
or the notes produced by Bannister [2.2]. These two publications were based on the fundamen-
tals ofpressure wave motion. Actually, one author, Jones [2.9], succeeded in solving the graphical
nmethod of characteristics by a computational procedure, and there are yet others who feel that
this chimera is worth pursuing.
In 1990, I began reexamining many of these computational procedures to improve the
quality of the design tool resulting from them, and found that, for one reason or another, all of
the available methods fell short of engine design needs. Put simply, a design model firmly
based on the motion of pressure waves is absolutely essential. At the same time, the computa-
tional model must be able to trace with great accuracy the rapidly changing gas properties in
the cylinder and in the exhaust pipe, where the exhaust gas from blowdown may be rapidly
followed by large amounts of fresh air, or fresh air and fuel, into the exhaust pipe and onward
256
through an exhaust catalyst or silencer. In addition, for marine engines, where the exhaust is
often internally water cooled, unless the model can readily incorporate the thermodynamic
effects of that water "injection," it will be less than useful as a design tool. Diesel engines, be
they two- or four-stroke cycle units also have gas properties in the exhaust system that vary
very considerably as a function of the load level.
Like the four-stroke engine, the two-stroke engine is heavily dependent on pressure wave
effects for the scavenging and charging processes and the model therefore must be based on
pressure wave transmission, propagation, and reflections. The theoretical model described
below must not be thought of as uniquely applicable to a two-stroke engine, for this is not the
case. Indeed, the many recent papers on inlet manifold "tuning," employed as a means of
modifying the volumetric efficiency-speed curve as a precursor to altering the torque curve,
indicates the growing interest that designers of four-stroke engines for production automobile
engines have in the prediction of unsteady gas motion. The designers of racing four-stroke
engines already knew that the "black art" that accompanied their thinking, as often expressed
to me, left much to be desired.
It should not be assumed that the lower-performance, four-stroke, spark-ignition engine is
immune to disadvantageous pressure wave effects. The common occurrence of blowback into
the inlet tract during the valve overlap period at the end of the exhaust stroke necessitates the
tracking ofvariable gas properties within that duct. If a simulation model cannot do this, then
a very inaccurate assessment of the ensuing air flow into the cylinder and of its trapped charge
purity will result.
A simulation model that describes the complexities of pressure wave motion in an engine
must meet one further requirement: Its principles must be understandable by the users, other-
wise the process of optimizing a given design will not only be long-winded and tedious, but
prone to decision-making on geometrical aspects of the engine that is illogically correct at one
engine speed and load, yet disastrously wrong at another. Such dichotomies must be thor-
oughly checked out by a designer who understands the fundamental principles on which the
model is based.
The basis of the GPB model employed here fulfills all of the criteria specified above.
2.18.1 The Basis of the GPB Computation Model
The unsteady gas flow process that is being modeled is illustrated in Fig. 2.20. This is the
same process that has been discussed throughout this chapter. The computational procedure is
somewhat similar to that found in other characteristics solutions. The pipe or pipes in the ducts
of the engine being modeled are divided into meshes of a given length, L. The pressure waves
propagating leftward and rightward are shown on Fig. 2.20, and at the instant that snapshot is
taken for a particular mesh labeled as J, the pressure values at the left end are PR and PL and the
pressure values at the right end of that mesh are pR1 and pLI. The gas in the mesh space has
properties of gas constant, R, and specific heats ratio, y, that are assumed to be known at any
instant in time. The reference density, po, and temperature, To, are also assumed known at any
instant. Because the diameter, d, and the volume, V, of the mesh space are a matter of
geometrical fact, the mass in the mesh space can be determined at that instant. The average
257
Pleftward Prightward
distancel
-mesh JI
Fig. 2.20 Meshing of the ductforpressure wave propagation.
pressure throughout the mesh space can be considered to be the mean of the superposition
pressures at each end ofthe mesh. The average superposition pressure amplitude ratio is given
by Xj:
XJ =
(XR + XL-1) + (XR1 + XLI -I) 2(2.18.1)
i
2
Consequently the average pressure, pj, density, pj, and temperature, TJ, are found from:
-=PXJ7 (2.18.2)
- POXG5 (2.18.3)
2=
TOX (2.18.4)
258

Emailing Design and Simulation of Four Stroke Engines

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Emailing Design and Simulation of Four Stroke Engines

Загружено:

Авторское право:

Доступные форматы

mv.

Design and Simulation of Two-Stroke Engines

Society of Automotive Engineers, Inc.

Copyright © 1999 Society of Automotive Engineers, Inc.

SAE Order No. R-186

*See Chapter 1, Ref. [1.9]

2.15 Reflection of Pressure Waves in Tapered Pipes .................... 236

Chapter 3 Discharge Coefficients of Flow within Four-Stroke Engines ........... 327

Chapter 5 Computer Modeling of Four-Stroke Engines ..................... 521

NAME SYMBOL UNIT (SI)

NAME SYMBOL UNIT (SI)

NAME SYMBOL UNIT (SI)

NAME SYMBOL UNIT (SI)

Introduction to the Four-Stroke Engine

Fig. 1.1 The four strokes of the four-stroke engine.

CR = maximum volume = V1d - Vsv + Vcv (1.1.1)

Plate 1.1 A cross section through a two-valve spark-ignition engine.

1.1.2 The Four Strokes

1.2 Cylinder Head Geometry of Typical Spark-Ignition Engines

Fig. 1.3 Valve actuation by a pushrod and rocker arm.

TlC BDC TDC BDC TPC

"CRANK EVO IVO EVC IVO

Fig. 1.4 Typical valve liftprofiles with respect to crank angle.

Fig. 1.5 A side-valve engine with two valves per cylinder.

Fig. 1.6 A four-valve pent-roof cylinder head.

1.2.4 Porting the Cylinder Head by Other Means

Plate 1.4 The 500cc Rudge Ulster motorcycle engine.

Plate 1. 6 The 350cc Barr and Stroud sleeve-valve motorcycle engine.

Fig. 1. 7A four-valve direct-injection (DI) diesel engine.

Plate 1. 7 A cutaway drawing ofa turbocharged direct-injection six-cylinder diesel engine.

1.3.3 The Three-Valve Indirect-Injection (IDI) Engine

Fig. 1.8 A three-valve indirect-injection (IDI) diesel engine.

1.4.2 Swept Volume

Vtsv = nVsv = n 4 dboLst (1.4.2)

1.4.3 Compression Ratio

VCV CR-i (1.4.3)

CRt = vts + VCV (1.4.4)

Fig. 1.9 The geometry of a connecting rod-crank mechanism.

Ftdc + Gtdc= (Lr + Lt)-D2 (1.4.5)

hence 0tdc= tan'(D (1.4.6)

hence Gtdc= Lct cosOtdc (1.4.7)

and Ftdc Lct cos Otdc + (Lcr + L2c)-t (1.4.8)

Fbdc = |(Lcr- ct)D2 (1.4.9)

hence Obdc tan-( (1.4.10)

and G bdc Lct (1.4.11)

The stroke of the piston is then given by:

Lst = Ftdc + Gtdc - Fbdc (1.4.12)

Ftdc Lcr + Lct

Fbdc = Lcr - Lct

Ht = (Ftdc + Gtdc) - (F + G) (1.4.14)

where G= Lct cosO (1.4.16)

and F= Lr-(E-D)2 (1.4.17)

Ht = (L2r + L Dt)D2 - cr - (L sin0 - D)2 - Lct cos 0 (1.4.18)

The angle of obliquity ofthe connecting rod, ¢, is given by:

F = (Ftdc + Gtdc-Ht )-G = k-G (1.4.20)

where a = -2DLct b = 2kLct c = k2 L_ct + D (1.4.22)

thenlet t = tana where a=-

consequently sine 2t and cos0= -

(c - b)t2 + (2a)t + (b + c) = 0 (1.4.23)

The solution for the crank angle, 0, becomes

0 =2 tanf1I -abc +b- )(1.4.24)

1.4.5 Numerical Data on Piston Position With Respect to CrankshaftAngle

Fig. 1.10 Sinusoidal and connecting rod-crank motion.

ERROR OF CONNECTING ROD-CRANK MOTION